Individual differences in event experiences and psychosocial factors as drivers for perceived linguistic change following occupational major life events

2.1 Setting the sociolinguistic stage: Austria as an empirical backdrop

Austria is a German-speaking country housing two main dialect areas, namely the Bavarian and Alemannic dialect regions (Wiesinger 1983). The Alemannic-speaking parts of Austria are characterized as diglossic and are geographically in the minority (i.e., the state of Vorarlberg and a few Tyrolean villages), whereas Bavarian dialects are spoken throughout the rest of Austria. Structurally, both Bavarian and Alemannic dialect varieties diverge from standard German varieties at the level of phonology, (morpho)syntax, and lexis.

Socio-stylistically, the use of Bavarian and Alemannic dialects has been found to be a function of macro-sociological (age, gender, interlocutor, contextual in/formality, etc.) and micro-situational (subject matter knowledge, topic, etc.) factors (e.g., Ender and Kaiser 2009), as well as identity-related issues (e.g., projection of meaning via socially indexed varieties). In addition, Austrians tend to have particularly positive attitudes toward dialect varieties – for example, in the Bavarian context, dialects are associated with social attractiveness, whereas standard German speakers are often perceived as more arrogant (though also more intelligent, educated, and professional) (Soukup 2009). Results from recent panel studies conducted in the Austro-Bavarian context suggest lifespan change in the form of increased dialect use in later life in comparison to during speakers’ economically active years (Bülow and Vergeiner 2021, Vergeiner et al. 2021), but remarkable stability as concerns their attitudes, specifically their dialect identity (Bülow et al. forthcoming). Whether and how occupational MLEs, as well as associated individual differences in event experiences and psychosocial factors, are related to changes in speakers’ linguistic repertoires, however, has not seen much empirical attention.

2.2 Occupational life-course transitions and individual differences as predictors of linguistic change

Sociolinguistic theory assumes that communally stable variants (or entire language varieties) can exhibit age-grading, that is, individual variation through time, but no resultant community-wide language change, and that these patterns may recur “at a particular age in successive generations” (Chambers 2008, 180). In this vein, Wagner (2012b, 180) attested that “nonstandard variants are associated with younger and (to a lesser extent) older speakers, but not middle-aged speakers, yielding a curvilinear frequency distribution.” This behavior has often been accounted for in terms of linguistic marketplace models (Sankoff and Laberge 1978), which propose that in being subjected to the pressures of the standard language-expectant workplace during economically active adulthood, speakers may register a retreat from vernacularity and increase their rates of standard language usage (Eckert 1997, Wagner 2012a). Upon disassociation from these social pressures, for instance, at or when approaching retirement, speakers may evince a revival of vernacular variants they had suppressed during their time in the workforce (Edwards 1992, Downes 1998, Buchstaller 2006, Buchstaller et al. 2017, Vergeiner et al. 2021, Pfenninger et al. 2023). For example, Mechler and Buchstaller’s (2019) study revealed two fundamentally different trajectories after retirement: (a) The aforementioned ‘tail’ for speakers past retirement: The three working class individuals illustrated the upward trajectory generally assumed in sociolinguistic theorizing (i.e., an increase in vernacularity), as opposed to (b) the continued standardizing effect of the sociolinguistic marketplace, as observed for the two socioeconomic risers who exhibited ongoing retrenchment toward the standard language. As Riverin-Coutlée and Harrington (2022, 42) remarked, however, the “idea of the career as a potentially influential external factor has not been empirically much explored,” despite the hypotheses that it may be a highly influential factor in lifespan sociolinguistic development (e.g., Sankoff and Laberge 1978, Eckert 1997, Buchstaller 2006, Wagner 2012a, Wagner 2012b, Vergeiner et al. 2021).

Most notably in the speech sciences has longitudinal evidence for career-related linguistic change surfaced. For example, Shapp et al.’s (2014) analysis of Ruth Bader Ginsburg’s speech indicated that her occupational transition from a lawyer to Justice at the Supreme Court temporally coincided with a reduction in her use of non-standard variants of New York City English. Homing in more specifically on career as a predictive variable, Riverin-Coutlée and Harrington (2022) found that advances in public French speaker Michaëlle Jean’s career path (e.g., her career becoming more international) correlated with changes in select phonetic features. That said, while these analyses indicate that MLEs may be change-inducing, the individual case studies are limited to ‘notable individuals’ (Sankoff 2018, 307) – for example, politicians, broadcasters, and actors among other public figures for whom digitally recorded speech data is in abundance. The case studies also represent ‘extreme case[s]’ (Riverin-Coutlée and Harrington 2022, 49), and the effects found are not likely to generalize to the average adult population (i.e., who are not highly public figures). Riverin-Coutlée and Harrington (2022) thus call for future work to conceive of designs that can model the influence of career on linguistic change across a larger number of participants. In addition, previous sociolinguistic research (Chambers 2008, Wagner 2012b, Mechler and Buchstaller 2019) suggests that the degree and directionality of lifespan change is likely to correlate with speakers’ educational and professional trajectories, satisfaction with those trajectories (including subjective wellbeing), and with the standardization pressures that result from their individual life choices. This underscores the need for designs that ideally consider not only the effects of macro-sociological individual differences (gender, education, etc.) on ensuing patterns of linguistic change, but also individual differences at the level of the event (e.g., how events are differentially experienced).

Life event researchers in psychology have emphasized that even the same MLE can be (and usually is) experienced differently across individuals (Rakhshani et al. 2022, Schwaba et al. 2023) – for example, some may perceive the same MLE as more positive, others as more negative. Given this, it is necessary to account for individual differences in event experiences (e.g., Schwaba et al. 2023), also referred to as event-related characteristics (Luhmann et al. 2021), the rationale being that it may not necessarily always be the event per se that induces change, but rather how the event is experienced by the individual. Similar hypotheses are also evident in sociolinguistics. For instance, as Riverin-Coutlée and Harrington (2022) postulated, significant changes in a speaker’s social status due to an MLE – what is regarded as a consequence-focused event-related characteristic in psychology (Luhmann et al. 2021) – may induce change in an individual’s linguistic repertoire (see also Bowie 2010). In other words, an approach focusing on event-related characteristics effectively allows us to separate the effects of the event in and of itself from how it is experienced and from the consequences one incurs because of it. In general, this should facilitate more nuanced insights into which aspects of significant life events drive linguistic change as opposed to the comparatively narrow inquiry about which events are associated with change.

In addition to classic sociolinguistic influences such as gender and educational attainment, individual differences in socioeconomic status and, by extension, occupation may also influence the degree of linguistic change an individual experiences following an occupational MLE. As Mechler and Buchstaller (2019), for instance, noted, individuals with significant marketplace pressures to employ standard language may carry these patterns across the threshold of retirement. Individual speaker differences in occupational complexity, for example, manual and communicative complexity of an individual’s primary occupation, may therefore emerge as important predictors for the degree and directionality of linguistic change as a function of a career-related life-course transition.

Finally, it seems feasible that psychological factors may impact on patterns of change (Andresen 2015, Steiner et al. 2023b) or otherwise moderate the relationship between event-related characteristics and linguistic outcomes. Specifically, Bowie (2010) noted the relevance of ‘coping’ mechanisms in processes of linguistic change, for instance how an individual linguistically copes in reaction to social pressures and events. In this vein, exploring psychological resilience, that is the ability to ‘bounce back’ and cope with stressful situations (Smith et al. 2008, Chmitorz et al. 2018), in interaction with event-related characteristics appears fruitful. Concretely, there seems to be a theoretical motivation for the interaction between psychological resilience and certain event-related characteristics: Since resilience resources have been hypothesized as relating to “successful engagement with difficult events and experiences” (Ryff and Singer 2003, 21), it seems plausible to explore whether, after experiencing an occupational MLE perceived as particularly negative and/or stressful, differences in reported linguistic change among individuals with varying levels of resilience may reflect some form of linguistic coping mechanisms, as Bowie (2010) questioned.

3.1 Research design

We integrate both quantitative and qualitative approaches by using an equal-status concurrent mixed-methods design (Johnson and Onwuegbuzie 2004), in which we place equal weight on insights generated via quantitative and qualitative approaches. The rationale for the mixed-methods approach was primarily that of expansion, that is “seeking to expand the breadth and range of research by using different methods for different inquiry components” (Johnson and Onwuegbuzie 2004, 22). Here, the quantitative analysis allows us to identify systematic sources of variation for differences in perceived linguistic change in relation to career-related significant life events. The qualitative approach gives participants an active voice to aid in advancing our current understanding of perceived MLE-related change, facilitating individualized insights into personal circumstances, attitudes, and idiosyncrasies that would otherwise be overlooked in a purely numerical analysis.

3.2 Participants

In this study, we analyze a subsample of participants from Wirtz and Pickl (forthcoming), specifically those who reported on occupational MLEs (first job, change of job, unemployment, and retirement). We removed from the full sample (a) individuals who did not report on an occupational MLE, and in addition to these (b) five individuals who did not provide information on occupational complexity, and also (c) one gender-diverse participant. The final sample pool thus comprised 154 Austrian participants between the ages of 18 and 76 (first job: n = 56; change of job: n = 55; unemployment: n = 3; retirement: n = 40), all of whom spoke German as a first language.

Figure 1 presents the sociodemographic characterization of the subsample, and Figure 2 presents the sociodemographic data disaggregated by MLE category. Given the crowdsourcing method (i.e., online experimental procedure) and thus the convenience sample, it was not possible to stratify for gender, age, or educational attainment, as is otherwise typical of sociolinguistic investigations. The distribution of participants across Austrian provinces is also uneven; this, however, does not pose a major problem given that most of Austria houses Bavarian dialect varieties and also because we focus on standard German and dialect rather than intermediate varieties, the existence of which is one of the main distinguishing characteristics between Austro-Bavarian and Austro-Alemannic dialect areas.

Figure 1

Sociodemographic information about the subsample.

Figure 2

Sociodemographic information disaggregated by MLE.

3.3 Tasks and procedure

The test battery comprised a single survey coded in LimeSurvey lasting approximately 10–20 min. We analyze here the results of six tasks, which we detail below.

Before beginning the survey, participants were informed that (a) participation was entirely voluntary, (b) the survey was anonymous and thus no information provided could be traced back to any individuals, (c) they could withdraw their consent and delete their answers at any time during the procedure, and (d) the data collected will be used purely for scientific purposes. Participants who agreed to these terms then began the survey.

3.3.2 Event-related characteristics

To capture participants’ perceptions of the identified MLE, we adopted Luhmann et al.’s (2021) Event Characteristics Questionnaire, which is a dimensional taxonomy of perceived characteristics of MLEs (i.e., individual differences in event experiences). Participants were asked about characteristics of the event that they believed had affected changes in their linguistic repertoire, specifically about the following event-related characteristics: (a) challenge (the degree of stressfulness and anxiety an individual associates with an event), (b) emotional significance (the degree of emotional impact the event had on an individual), and (c) social status changes (the extent of negative changes in an individual’s social status). Responses were provided on 100-point slider scales from ‘does not apply at all’ to ‘applies completely’. These event-related characteristics were chosen because the variable use of standard German and dialect in the Austrian context has been related to differences in emotionality (e.g., Huesmann 1998, Kroisenbrunner 2015) and social status (e.g., Steinegger 1998, Ender and Kaiser 2009), and thus, the extent to which an event is perceived as inducing changes in these characteristics may correlate with shifts in the sociolinguistic repertoire as well.

In line with Luhmann et al.’s (2021) extensive item analysis procedures, Cronbach’s alpha indicated acceptable internal consistency (computed on the basis of the full sample, i.e., N = 701; challenge: α = 0.85; emotional significance: α = 0.81; social status changes: α = 0.89). Figure 3 presents the descriptive statistics for the event-related characteristics disaggregated by MLE category (note that the individual differences for unemployment are not shown, as this MLE is not investigated in the quantitative analysis due to the small sample size).

Figure 3

Descriptive statistics of the event-related characteristics disaggregated by MLE.

3.3.4 Psychosocial variables

To capture participants’ proficiency in standard German and dialect (i.e., varietal proficiency), responses to four items on informants’ productive (speaking) and receptive (listening) language proficiency on 100-point slider scales were collected, based on the following item adaptations: ‘How well do you [speak/understand] [standard German/dialect]?’ This variable was included in the subsequent models given that Austrians’ varietal proficiency has been shown to vary (Ender and Kaiser 2009); thus, the degree of MLE-related change in their cross-contextual use of a certain variety may also be contingent on their proficiency in the respective variety.

Participants also completed the German version of the Brief Resilience Scale (Chmitorz et al. 2018), which is a short six-item measure that assesses their ability to ‘bounce back’ from stress and overcome stressful, particularly negative situations. In line with Chmitorz et al.’s (2018) validation study, the items also show acceptable internal consistency (computed on the basis of the full sample, i.e., N = 701; α = 0.84). Responses were provided on 100-point slider scales from ‘strongly disagree’ to ‘strongly agree’.

Finally, participants completed a questionnaire addressing their occupation-related communicative and manual complexity. The items for these scales were based on Zacher and Frese’s (2011) scale items gauging general job complexity, which refers to “the extent to which the tasks on a job are complex and difficult to perform … work that involves complex tasks requires the use of numerous high-level skills and is more mentally demanding and challenging” (Morgeson and Humphrey 2006, 1323). In line with this definition, our scales reflect the degree to which an individual’s job requires an individual to regularly communicate with others (e.g., ‘In your (if retired, former) job: How often do you need to talk to other people to complete your work tasks?’) and/or complete manual handling tasks (e.g., ‘In your (if retired, former) job: How often do you have to carry out manual work (possibly with different materials)?’). The two scales show acceptable internal consistency (computed on the basis of the individuals taken from the full sample who reported being or having been active in the workforce N = 600; occupation-related communicative complexity, three items: α = 0.65, though this is indeed on the lower end of accepted values; occupation-related manual complexity, three items: α = 0.86).

Figure 4 presents the descriptive statistics for the psychosocial variables disaggregated by MLE category.

Figure 4

Descriptive statistics of the psychosocial variables disaggregated by MLE.

3.4 Data analysis

4.1 Event-related and psychosocial predictors of occupational MLE-related linguistic change

Figures 6 (first job), 7 (change of job), and 8 (retirement) present the visual model summaries from the Bayesian models assessing the influence of event-related characteristics and psychosocial and demographic variables on reported patterns of perceived MLE-related linguistic change in the form of quantile dot plots (without outliers, see OSF for the model summaries including the outliers). These visualize the posterior probability distribution of the predictor variable’s effect size. Each dot in the figure represents a 1% likelihood of a given value. Effects that overlap with the ROPE are shaded gray, indicating non-sufficient evidence for an effect. Green and red shadings show a significant effect with a positive and negative directionality, respectively.

Figure 6

Visual model summaries for the effects of individual differences on perceived linguistic change in relation to entry into the workforce (n = 47 after outlier removal; men = 10, women = 37; higher education = 37; no higher education = 10). Note: The purple-colored bars show the 50%, 80%, and the 95% HDIs (from darker to lighter). The black point displays the posterior mean, and the thin and thick black bars visualize the 98% and 66% HDIs, respectively. The red-shaded area is the ROPE set at ±0.10.

Figure 7

Visual model summaries for the effects of individual differences on perceived linguistic change in relation to change of job (n = 45 after outlier removal; men = 17, women = 28; higher education = 26; no higher education = 19).

Figure 8

Visual model summaries for the effects of individual differences on perceived linguistic change in relation to retirement (n = 33 after outlier removal; men = 13, women = 20; higher education = 28; no higher education = 5).

Upon entry into the workforce, differences in perceived changes in cross-contextual use of and attitudes toward standard German are moderated by proficiency in standard German, with more proficient individuals evincing a more pronounced self-reported retrenchment toward standard language (Figure 9a and c). We also found that individuals reporting that their occupations required them to perform at least some manual tasks were more resistant to changes in patterns of accommodation (Figure 9b), though this subsample evinced overall low rates of manual job complexity (see Figure 4).

Figure 9

Conditional effects of significant predictors for perceived linguistic change following entry into the workforce. Note: We superimposed the raw data onto the conditional effects plots, with each point representing a participant.

Event-related characteristics significantly predicted perceived MLE-related change exclusively among individuals who changed jobs. Participants who perceived their change in job to be particularly damaging to their social status reported a retreat from standard language (Figure 10a) and an uptake in cross-contextual dialect use. Conversely, individuals who perceived the change in their career trajectory to be especially stressful reported a retrenchment toward standard language (Figure 10b).

Figure 10

Conditional effects of significant predictors for perceived linguistic change following a change of job.

In the unadjusted models (i.e., including the outliers), we found two interaction effects between psychological resilience and event-related characteristics (see the visual summaries for the unadjusted models on OSF), but these effects did not withstand outlier removal. We return to this point in the discussion.

Interestingly, none of the individual difference variables significantly predicted differential patterns of perceived linguistic change following retirement.

4.2 Beliefs about individual drivers for linguistic change

We now draw on the qualitative data collected in the open-ended questionnaire items (see the Codebook on OSF) in order to expand on the previous findings from another angle. These data were subject to a reflexive TA, which reflects and honors the subjectivity of participants’ lived experiences, views, and perspectives as concerns which aspects of an MLE were particularly salient in motivating linguistic change.

Across the occupational MLEs, facets of the linguistic market manifested in participant responses. For example, standardization pressures as drivers for patterns of retrenchment toward standard German (i.e., ‘High German’) were mentioned by informants in the workforce (our translations):

  1. When I started working, I was required to speak in High German (P369_W_NÖ_First_Job).

  2. Starting work meant that I often had to speak standard German instead of colloquial language (P608_W_SBG_First_Job).

  3. Due to my job, it became necessary to speak more High German (P543_M_SBG_Change_Job).

That said, not every workplace is equally standard expectant. A plethora of work-related aspects such as job requirements and social networks will affect speakers’ linguistic conduct. For instance, a multicultural team, or at least a team with less affinity for dialects, may give rise to considerable linguistic accommodation necessities to facilitate communication among coworkers:

  4. Multicultural team, therefore, avoidance of dialect expressions (P562_W_SBG_Change_Job).

  5. Supra-regional work requires accommodation (P1366_M_TRL_Change_Job).

  6. Nobody at work would understand my dialect (P1261_M_WIEN_First_Job).

By contrast, a less standard language-conformant workspace, and especially ones in which contact with dialect is the norm rather than the exception, may be conducive to an increase in vernacularity:

  7. Administrative work with contact to parties from all over Austria (lots of telephone contact with callers from Vorarlberg to Burgenland every day, […]), one colleague with a strong Eastern Austrian dialect, another from the Rhineland (P1383_M_WIEN_First_Job).

  8. Since 2016, I have been working with people from all walks of life and different regions of Salzburg. In conversations, I realized how many different dialects are still alive in our country. As a result, I also specifically use my dialect where it is appropriate and also like to play with it (P568_W_SBG_Change_Job)

  9. Changing jobs to a company where a strong regional dialect is spoken (P1025_M_TRL_Change_Job).

10. Change of job to a much more down-to-earth industry. Fewer colleagues with a high affinity for High German, more conversations with people from rural areas (P1349_W_TRL_Change_Job).

Advances in the career path and the associated shifts in social status, communicative necessities, social networks, etc., may also play a role in shaping the (socio)linguistic repertoire and in strengthening pressures to employ standard language:

11. As a supervisor, I had to pay more attention to my choice of words and their meaning (P636_W_SBG_Change_Job).

12. Since I’ve been lecturing all across Austria, I almost only use High German. This was probably reinforced by my online seminars, too (P449_W_WIEN_Change_Job).

13. Due to my position in a working environment with many highly educated people, I am forced to speak more High German (P706_W_WIEN_Change_Job).

Importantly, shifts in social status need not always occur in an upward direction in one’s career trajectory, and perceived ‘downward’ social mobility may coincide with a potentially less standard-expectant workplace and thus facilitate (a revival of) vernacularity:

14. Most of my [previous] colleagues were from the city and spoke High German, while my colleagues at work come from rural areas, speak dialect, and have a different educational background (P549_W_SBG_Change_Job).

Respondents reporting on the impact of the retirement threshold noted a clear drop in standardization pressures associated with economically active adulthood, linking this with a retreat from the standard language toward dialect varieties. In addition, retired individuals clearly viewed this relaxation of social and work-related pressures to employ ‘pretty language’ (i.e., standard language) in a positive light:

15. The use of High German in dealing with customers has gone away (P415_M_OÖ_Retirement).

16. As a teacher, I had to speak standard language. Now I use my Pongau dialect as often as possible (P736_M_SBG_Retirement).

17. Spoke a lot of standard language at work, but no longer a compelling need in retirement (P1315_W_OÖ_Retirement).

18. Due to my retirement, I am no longer “forced” to speak High German or Standard German with colleagues and employees, which means I automatically speak a lot more dialect (P129_W_SBG_Retirement).

19. I’m no longer required to use the standard language for professional reasons (P221_W_SBG_Retirement).

20. There’s no longer any need to speak standard German (P223_W_SBG_Retirement).

21. It is no longer necessary to speak beautiful ‘written German’ (P535_W_SBG_Retirement).

Interestingly, this perceived retreat from standard language in reaction to weakening standardization pressures is not only evident for retired individuals in later life, but may manifest in general upon disassociation from a standard-expectant workplace, as one of the participants reporting on the effects of unemployment stated:

22. Unemployment, […] loss of colleagues with whom you spoke High German (P1395_M_WIEN_Unemployed).

Finally, participants also expressed socio-affective aspects that may moderate the relationship between career-related life-course transitions and individual differences in linguistic change. One particularly recurrent theme was dialect-related pride experienced during or because of a career-related event, which may account for perceived MLE-related increases in vernacularity and a strengthened sense of dialect identity:

23. In the beginning, I tried to keep work colleagues from noticing that I came from the “countryside,” but after a few years the opposite was the case, I was and am happy and proud that I can speak dialect, understand dialect, that you can tell where I come from (P536_W_SBG_First_Job).

24. Many colleagues from Germany. Greater focus on my own dialect. Became more important to me. The local ‘color’ [i.e., dialect, MW/SP/SP] is very pleasant compared to the dry, hard High German in everyday working life (P520_W_SBG_First_Job).

25. [Because I’m a] nurse in a retirement home: Old/forgotten words in dialect are reintegrated into everyday speech. You learn to appreciate your own dialect more and want to pass it on/keep it alive (P262_W_SBG_First_Job).

26. As a native of the Innviertel region, I found High German […] to be unnatural, and I am happy to now speak more cultivated dialect than in my childhood and youth, to be understood and to be accepted linguistically (P229_W_OÖ_Retirement).

27. I’m using my “learned” dialect [i.e., the first-acquired dialect, MW/SP/SP] again and I’m glad that others do this too and appreciate it (P633_M_SBG_Retirement).

5.1 RQ1: Effects of quantitatively captured individual differences on reported linguistic change

Life event researchers in psychology recognize that even the same MLE can be and typically is experienced differently across individuals (Rakhshani et al. 2022, Schwaba et al. 2023), and recent sociolinguistic evidence points in a similar direction (see, e.g., Mechler and Buchstaller’s [2019] analysis of different linguistic trajectories when approaching or after retirement). In order to deal with this, we explored individual differences in event experiences (e.g., Luhmann et al. 2021, Schwaba et al. 2023), an approach which effectively allowed us to separate the effects of the event in and of itself from how it was experienced by the participants.

We found that negative changes in social status incurred especially after a change in job correlated with a perceived retreat away from the standard language. In other words, individuals who perceived a change in their career trajectory to induce a shift down the social ladder may draw increasingly on nonstandard variation, a finding that is in line with previous empirical evidence illustrating that (perceived) social mobility may “counteract declining plasticity” (Sankoff 2018, 299) and thus turn out to be a powerful driver for patterns of (perceived) linguistic change (e.g., Beaman 2021, Riverin-Coutlée and Harrington 2022).

At the same time, we identified nuanced interaction effects between event-related characteristics and resilience resources in the unadjusted models, but these effects did not withstand outlier removal (the directionality of the effects in both modeling procedures did, however, remain the same). In the unadjusted models, we found, for instance, that in the face of a high-stress job change, more resilient individuals reported more positive attitudes toward standard German. This may be a first indication of what Bowie (2010) considers ‘linguistic coping’. That is, particularly resilient individuals may adjust their sociolinguistic repertoire more noticeably in order to better navigate high-pressure professional contexts – ones which likely go hand in hand with high workplace pressures to employ standard language. This insight, however, remains speculative, and future research with a focus on the interaction between psychological resources (e.g., resilience) and significant life-course transitions may consider addressing this issue further to determine whether the effect becomes more or less pronounced with more data.

Interestingly, entry into economically active adulthood was not reliably moderated by any event-related characteristics. Rather, our results indicate that individual differences primarily in proficiency are more predictive of individual-level perceived linguistic change. For example, individuals more proficient in standard German reported a stronger retrenchment and more positive attitudes toward standard language. While the regression analysis inherently indicates a predictor-outcome relationship, this effect is likely more circular in nature. That is, speakers who have a higher prior proficiency for the standard language are able to undergo more pronounced retrenchment, and those who have undergone more retrenchment will likely evince a higher proficiency in standard German. A similar circular pattern can be argued for their attitudes toward standard language as well. In addition, we found that individuals who reported that their job required at least some manual tasks were more stable in their accommodation practices following the MLE entry into the workforce. That said, the overall manual complexity scores were low in this subsample, which raises the question as to whether it is truly manual job complexity associated with the more stable patterns of accommodation, or whether related job components not captured in our two occupation-related scales may be manifesting here.

5.2 RQ2: Qualitative insights on between- and within-event variation in patterns of linguistic change

The TA also illustrated that a single significant life event need not impact in the same way or for the same reasons on the sociolinguistic repertoire. Let us consider the MLE entry into the workforce. The descriptive data evinced a mean increase in cross-contextual standard German use and a moderate reduction in reported rates of dialect. This general trend comes as little surprise, especially given the extensive sociolinguistic theorizing that entering economically active adulthood may coincide with linguistic retrenchment toward standard language and, at the same time, cause a retreat from vernacularity (e.g., Eckert 1997, Wagner 2012a), likely owing to the standardization pressures of the linguistic marketplace (Sankoff and Laberge 1978, see also Labov 1966, Wolfram 1969, Trudgill 1974, Bowie and Yaeger-Dror 2015). Indeed, aspects of the qualitative analysis confirm that marketplace influences are at play for many individuals, driving retrenchment toward standard language and necessitating interpersonal accommodation. However, there do seem to be several variables influential enough to mitigate the effects of typical normative forces at play during an individual’s career-active years and even nudge speakers out of the typical shift toward conservative norms. For example, job tasks placing little pressure on speakers to employ standard German and colleagues with an affinity for vernacularity contribute to a generally less standard language-conformant workplace. In addition, a conscious assertion of identity and personal pride in one’s own dialect, even in the face of high-pressure workplace norms for standard language, may give rise to higher (perceived) rates of dialectality. These findings bear parallels to Wagner’s (2012b) analyses scrutinizing the transition from adolescence into adulthood (specifically, into higher education): She found that individuals with new college networks and at prestigious universities significantly reduced their use of nonstandard variants, whereas those who retained strong neighborhood ties remained stable in their use of the nonstandard variants. Like our results, this underscores the critical role of social networks, feelings of identity and belonging, and in general of event-related individual differences (e.g., the degree of standardization expectations in an MLE-resultant context) in how perceived changes in linguistic behavior manifest in relation to a significant life-course transition.

Whereas we found different trajectories of perceived change for individuals reporting on the linguistic relevance of entry into the workforce and a change in job, retirement evinced a remarkably stable interindividual directionality of perceived change patterns, at least in the productive domain. Apart from a few select outliers, retirement was strongly associated with a self-reported increase in dialectality and retreat away from standard language, a finding which is congruent with existing panel studies on Austrian speakers (e.g., Bülow and Vergeiner 2021, Vergeiner et al. 2021). In line with previous sociolinguistic theorizing (Edwards 1992, Downes 1998, Cheshire 2005, Buchstaller 2006, Vergeiner et al. 2021), our results underscore that this perceived linguistic behavior relates to weakening societal and workplace standardization pressures, the disassociation from which was perceived very positively. The affectively positive view toward relaxed conservatism generally corresponds to the reported changes in both the productive and affective-attitudinal aspects of the sociolinguistic repertoire and harmonizes nicely with Mechler and Buchstaller’s (2019) analysis illustrating that some speakers may proudly avoid social norms relating to the use of standard language in later life.

5.3 Limitations and future directions

To begin, one major limitation of this study is our focus only on reported measures of linguistic change in relation to occupational MLEs in relatively well-educated adults. Indeed, previous work from psychology focusing on personality development emphasizes that retrospectively judged change correlates moderately to strongly with measured change (Schwaba et al. 2023), but whether the same holds for retrospective judgments of linguistic change is an outstanding question. Future work focusing on the effects of significant life events would thus do well to collect speech and attitudinal data both in real time and retrospectively in order to correlate longitudinal trajectories with retrospective self-reports of linguistic change – possibly with participants that do not meet the WEIRD criteria (Western, Educated, Industrialized, Rich, and Democratic).

An additional limitation of the present method was that the participants were required to identify one MLE which they believed had the most pronounced impact on their sociolinguistic repertoire. This obscures the fact that linguistic behavior (and, by extension, perceptions thereof) is shaped by a multitude of forces, including, for instance, sequences of MLEs, day-to-day experiences, and so forth. Given this, the perceived effects of a single MLE may be overwhelmed by the cumulative effects of other factors. In merging both quantitative and qualitative analyses, the latter offering richer insights into participants’ contextual background which would have been lost in a purely numerical analysis, we attempted to mitigate this bias, but future analyses should strive to better ascertain the interactive effects of MLEs on changes in the (socio-)linguistic repertoire.

In addition, this study is the first to our knowledge to have looked at the linguistic relevance of unemployment. Despite the small sample size for this event (n = 3), the qualitative insights in combination with the descriptive data suggest that, at the least for one individual, the disassociation from the linguistic marketplace owing to unemployment curtailed the cross-contextual use of standard language, but at the same time gave rise to more positive attitudes toward standard language. Future research may consider investigating the linguistic relevance of unemployment on patterns of linguistic change, and how (the short-term disconnect from) linguistic marketplace pressures modulate individual-level linguistic change.

Finally, we believe that the design of the current study and analysis is noteworthy in sociolinguistic lines of inquiry. As Wagner (2012b, 182) puts it, “large studies [cannot] explain the individual differences without more detailed ethnographic information, whereas the smaller studies are unable to untangle the effects of social and/or stylistic factors.” By meaningfully integrating different investigative components, we were able to explore both systematic individual difference predictors of MLE-related linguistic change without neglecting individualized insights into personal circumstances, attitudes, and idiosyncrasies. In other words, complementary methods of investigation can facilitate deeper engagement with the multilevel and multivariate analysis of complex issues involved in sociolinguistic development and allow us to “obtain data about both the individual and broader societal context and bring out the best of the qualitative and the quantitative paradigms while also compensating for their weaknesses” (Pfenninger and Singleton 2017, 5).