On the relationship between complexity of primary occupation and L2 varietal behavior in adult migrants in Austria

1 Introduction

Occupational status is perhaps one of the most pervasive factors during adulthood. As Eckert (1997: 167) put it, work-related benchmarks in adulthood, for example a person’s first full-time employment, promotions, job loss, career changes, retirement, and so on, are “life experiences that give age meaning” and can critically influence the way an individual draws on language (variation) as a communicative tool. When it comes to second language (L2) learners, recent decades have made increasingly apparent (e.g., George 2014; Howard et al. 2013; Regan et al. 2009) that exposure to the target-language community is one of the most crucial factors in acquiring sociolinguistic competence, that is, “the capacity to recognize and produce socially appropriate speech in context” (Lyster 1994: 263; for an overview, see Geeslin and Long 2014; Howard et al. 2013; Regan et al. 2009). That said, the relationship between exposure and, for instance, production of variable structures is “neither causal nor categorical across all learners” (Geeslin 2018: 557; see also Geeslin and Gudmestad 2008). This necessitates expanding our purview to encompass factors complexly interwoven with the intensity with which migrants are exposed to sociolinguistic variation, such as their occupational status and the complexity thereof, and how these factors impact on differential outcomes in sociolinguistic development.

Specifically, we explore the effect of a factor not yet examined under a sociolinguistic lens, namely “occupational complexity”. Occupational complexity is a measure that allows us to more systematically tease apart different complexity characteristics of occupations. In the Dictionary of occupational titles (1977; henceforth DOT), occupations were rated based on observations by job analysts, and the resultant occupational information classifies jobs based on their complexity with, for example, “people” and “things”. Higher “complexity with people” refers to occupations with more monitoring and supervising, necessarily implying a more extensive need for communicative competence, whereas higher “complexity with things” is associated with occupations in which more handling and physical precision work is involved (Smart et al. 2014; Sörman et al. 2019). Importantly, the novelty and necessity in considering the independent variable occupational complexity is that it allows us to circumvent using arbitrary categories or thresholds as factors that function only as proxies for occupation, such as socioeconomic status or education. In other words, we are able to more directly gauge how different complexity dimensions (e.g., relating to communicative or physical components) of an individual’s occupation impact on patterns of language variation.

Thus, the aim of this paper is to explore how occupational complexity factors (with “people” and “things”) of primary occupation in the naturalistic (i.e., target-language) environment impact on migrants’ interpersonal varietal behavior. We analyze migrants’ differential use of standard German, Austro-Bavarian dialect,^[1] and mixture varieties in interaction with a standard German-speaking and a dialect-speaking interlocutor, which serves as a measure of learners’ sociolinguistic competence (Ender 2022; Wirtz 2022). To our knowledge, no studies in variationist second language acquisition (SLA) exist that attempt to (i) operationalize occupational complexity in such a multifactorial way and (ii) elucidate how the complexity of migrants’ primary occupation in the target-language environment relates to their varietal behavior. Thus, not only thematically, but also methodologically, this study paves new ground with respect to how we can operationally define and capture the complexity of occupational status and more fruitfully integrate it into both population- and individual-level analyses. Our results should therefore be of interest not only to (variationist) SLA scholars, but to sociolinguists more generally, as this study sets out a new direction for research questions investigating how career-related differences impact on language variation and use.

2 Theoretical background

At the broadest level, L2 learners’ exposure to and experience with the L2 target culture have been positioned as key stepping stones in acquiring sociolinguistic variation (George 2014; Kinginger 2008; Regan et al. 2009; Ringer-Hilfinger 2012). In the German-speaking world (specifically, in the Swiss-Alemannic context), Ender (2015, 2017, 2021, 2022 was the first to explore (influencing factors of) the differential use of standard versus nonstandard German varieties in interaction with standard German- and Swiss dialect-speaking interlocutors by adult migrants with L2 German. Of her 20 participants, 15 speakers evinced a relatively invariable preference for the standard variety (5 users), the local dialect (2 users), or mixed speech (8 users). Only five participants shifted varieties in an interlocutor-dependent way. Among other (psycho)social factors, Ender (2017, 2021 underscored the crucial role of the intensity with which migrants agentively expose themselves and/or are socially or occupationally exposed to the L2 target culture in occasioning how they differently integrate sociolinguistically functional varieties into their multi-varietal repertoires.

Intertwined with the intensity and type of sociolinguistic variation migrants are exposed to is their occupational status (Ender 2022; Prediger 2016) and, by extension, the communicative, technical, or physical activities required therein. The Austro-Bavarian context houses a continuum of language varieties ranging from traditional Austro-Bavarian dialects to standard German varieties, all of which coexist in everyday life (Ender and Kaiser 2009) and are distinguishable on structural and socio-situational fronts (for examples, see Kaiser 2022). Against this backdrop, it has been repeatedly noted (e.g., Vergeiner 2019, 2021) that occupations involving more manual handing are associated with higher Austro-Bavarian dialect usage, whereas professions requiring more mobility, higher formality, or public addresses necessitate standard German varieties. In light of such career-related (socio)linguistic variation, potentially beneficial effects of primary occupation in the naturalistic sphere on migrants’ acquisition of sociolinguistic competence may be of special interest, especially when considering the amount of time many individuals spend in their workplace.

That said, capturing the role of occupational status on migrants’ varietal behavior is easier said than done, and there has long been a dissatisfaction in sociolinguistics concerning the “failure to explain the quantitative relationships between linguistic variables and socioeconomic characteristics” (Forrest and Dodsworth 2016: 31; italics in original). Typically, participant occupations are subjected to some sort of (mono-factorial) categorization. For example, Prediger (2016) classified the occupational status of his L2 German participants on a five-point scale based on the degree of manual labor required. Though he indeed found that participants with higher manual requirements produced more informal dialectal variants, such a mono-factorial categorization drastically underestimates the complexity of the variable “occupational status”.

Relatedly, there is a clear dearth of research from both sociolinguistic and SLA perspectives that attempts to directly address the effect of occupation on patterns of language variation. Most work takes socioeconomic status as a sort of proxy variable. For example, early sociolinguistic research conceptualized class as a 10-category linear scale (Warner and Lunt 1942; Warner et al. 1949), which aggregated education level, family income, and occupational rank. The goal was that such an operationalization would multifactorially tap into several different dimensions of socioeconomic status (e.g., Labov 2001). Indeed, other class-related or socioeconomic factors do often correlate with occupational rank or status (Dodsworth 2009) – however, such macro-definitions of socioeconomic class or status are not necessarily able to capture within-group variation. For example, Chambers (2008) noted that among an occupational group called “professionals”, certain individuals practice careers in which adherence to linguistic norms is less important (e.g., engineers) while for others adhering to linguistic norms is critical (e.g., teachers). Decomposing different components of an individual’s occupation, rather than aggregating occupations according to an oftentimes arbitrary cutoff point in relation to socioeconomic status, therefore becomes a necessity. What is more, because of the dense bandwidth in multifactorial communicative, technical, or physical skills necessary for each occupation, different components of an individual’s occupation are likely to influence the degree of contextual diversity and, by extension, the locality- and population-level sociolinguistic variation to which one is exposed. Thus, adhering to operational definitions directly of occupation on multifactorial and continuous scales allows us to better capture the granularity in how differences in physical and communicative demands in individuals’ primary occupation relate to patterns of (person-level) variation.

In light of this, we attempt to capture the role of two job properties by drawing on the work characteristic “occupational complexity” (see Section 1), a multifactorial measure that defines occupations with respect to complexity with (i) data, (ii) people, and (iii) things, though we focus exclusively on (ii) and (iii) as these are the most relevant for our hypotheses (see Sections 3 and 4.2). Relating occupational complexity to variationist SLA, it seems plausible that complexity of migrants’ primary occupation may – for reasons related to identity construction, such as the assertion of group membership – influence the way they engage in sociolinguistic variation.

3 Aim and scope

This study approaches the role of occupational complexity on migrants’ interpersonal varietal behavior from two main perspectives: (1) investigating the relative contribution of individual differences in complexity with people and things and (2) seeking to gain (qualitative) insight into how L2 learners’ deviations from group-level patterns relate to the complexity of their primary occupation in the naturalistic environment.

The following research questions are addressed:

1. What is the strength of the association between occupational complexity (with people and things) and differential outcomes in adult L2 German learners’ interpersonal varietal behavior?

2. How are migrants who deviate from group-level estimates characterized in terms of the complexity of their primary occupation?

Because it has been asserted that individual-level factors such as occupation modulate varietal behavior in the Austro-Bavarian context (Scheuringer 1990; Steinegger 1998; Vergeiner 2021), and that first language (L1) speakers in manual labor careers tend towards dialect varieties, we expect migrants with higher occupational complexity with things to evince higher rates of dialect usage. As for how occupational complexity with people plays into migrants’ patterns of varietal behavior, this remains an open question.

Given the limits of this short article, this study exclusively explores the impact of the macro-contextual variable “occupational complexity”. These findings should be considered in conjunction with the analyses of the same learners presented in Wirtz and Pfenninger (2023a) and Wirtz et al. (2024), in which the goal was to elucidate how L2 sociolinguistic development is quantitatively and qualitatively mediated by a complex constellation of psychosocial individual differences.

4 Materials and methods

5 Results

Table 1 presents the descriptive statistics of the relative frequency of the 36 participants’ interpersonal varietal behavior in interaction with the standard German-speaking and the dialect-speaking interlocutors. These 36 participants produced 3,035 clausal units in total (1,558 with the dialect-speaking interlocutor and 1,477 with the standard German-speaking interlocutor; see Wirtz et al. [2024]).

5.1 Group-level analysis

First, we attempt to establish whether there were any population-level effects of occupational complexity on participants’ differential use of standard German, Austro-Bavarian dialect, and mixture varieties in interaction with the dialect-speaking (Model 1) and the standard German-speaking (Model 2) interlocutors. Figure 1 provides a visual model summary of the posterior distribution of the effect sizes for the two models. As already noted, Bayesian models result in an entire distribution for each parameter. The quantile dot plots in Figure 1 visualize the posterior probability distribution for each effect size as 100 equally likely points. The more points stacked on top of each other, the more probable a value in the distribution is for the specific effect; and the more dots (representing the distribution) that fall outside of the ROPE (i.e., the red area, set at ±0.18), the more compelling the evidence for the directionality of the effect.

Figure 1:

Visual model summary (in log-odds) of interpersonal varietal behavior as a function of occupational complexity. “Dialect” and “mixture” at the top of each plot indicate the log-odds ratio probability distribution of producing a dialect or mixture variety, respectively, in comparison to the standard German variety. The purple-colored bars depict (from darker to lighter) the 50 %, 80 %, and 95 % highest density intervals. The black points with bars represent the means (the points), the 98 % (thin bars), and 66 % (thicker bars) highest density intervals.

In Model 1, neither complexity with people nor with things was a robust predictor of participants’ increased use of dialect and mixture varieties as opposed to standard German. At most, there was a slight positive relationship between complexity with things and participants’ inter-individual use of dialect and mixture varieties.

Model 2 revealed that, when it comes to the relationship between complexity with things and migrants’ inter-individual use of dialect, the 95 % HDI included strictly positive values, indicating that the effect is slightly more robust compared to the other effects. While we must attribute uncertainty to the effect, the model suggests a comparatively strong tendency for migrants to produce dialect as opposed to standard German with increasing occupational complexity with things, that is, in occupations requiring more handling and physical precision work.

5.2 Person-centered insights

In order to complement the group-level analysis with person-centered insights, we specified two additional intercept-only Bayesian multinomial models including by-participant and by-VR set random intercepts. Here, our goal was not to identify population-level effects, but rather to home in on the individual. This analysis is of particular methodological importance, as it allows us to more closely observe individual job-related proprieties over which the overarching occupational complexity measure perhaps aggregates. What is more, such an individual-level view of the participants’ varietal behavior in relation to their occupational complexity metrics may allow us to capture cultural specificities of these migrants’ primary careers that are necessarily neglected by the DOT classifications (e.g., because they were determined based on occupations in the United States). Following Wirtz and Pfenninger (2023b), we targeted specifically those individuals who deviated most notably from the group-level mean in terms of their addressee-relational varietal behavior, specifically the “outliers”, who present particularly beneficial case studies to paint supplementary pictures of “individual stances against the backdrop of … overall trends” (Regan 2013: 29; see also Ellis and Larsen-Freeman 2006; Kanwit 2019, 2022).

Figures 2 and 3 plot 100 pulls from the posterior distributions of individual intercepts (i.e., deviation from the group-level mean in terms of their dialect and mixture production patterns) along with the mean value for each participant; additionally, the color indicates the participant’s occupational complexity level. The higher above the dotted line (representing the mean) participants are, the more they produced the respective variety shown in the plot in relation to the other participants. Visual inspection revealed that among some of the most notable outliers who deviated above the group-level mean in interaction with both virtual interlocutors (i.e., who produced more dialectal and mixture speech) were the participants outlined in Table 2.

Figure 2:

Individual-level deviation from the mean in interaction with the dialect-speaking interlocutor.

Figure 3:

Individual-level deviation from the mean in interaction with the standard German-speaking interlocutor.

Table 2:

Occupational traits and complexity of the “outliers”.

Participant ID	Occupation	Occupational complexity (people/things)
10_j_30_c	Mobile hydraulic technician	0/6
15_p_30_c	Software engineer	2/5
16_a_25_c	Musician	4/6
21_l_30_c	Teacher	6/0
22_c_24_c	Salesperson	3/0
24_j_34_c	PhD candidate	2/0
29_v_28_c	Medical student	2/0

As Table 2 and Figures 2 and 3 show, participants with a high occupational complexity with things (i.e., 10_j_30_c, 15_p_30_c, and 16_a_25_c) tended to deviate above the group-level means with respect to their dialect and mixture varietal use, underscoring the aggregate model from the previous section. That said, there were also clear instances in which participants from nonmanual occupations engaged in dialect and mixture varietal behavior. For example, participant 22_c_24_c, a salesperson, noted that “at work […] we only speak in dialect”. Along similar lines, participant 29_v_28_c, a medical student, also evinced dialect and mixture varietal behavior above the group-level mean. Arguably, this may reflect her chosen occupation in the medical field, in which she is likely confronted with dialect-speaking patients (e.g., during internships), inherently necessitating a high level of dialect proficiency, and underscored by her statement that “I have to use it [i.e., dialect] or it [i.e., life] just doesn’t work”. A similar argument can be made for participants 21_l_30_c and 24_j_34_c, a teacher and a PhD candidate: in light of the comparatively accepted standing of dialect and mixture varieties in school and university contexts (Buchner et al. 2022; Vergeiner et al. 2019, 2021), navigating the varietal waters in such occupations requires proficiency not only in standard German, but in Austro-Bavarian dialect as well.

6 Discussion and conclusion

The goal of the current study was to address the effect of occupational complexity with people and things on L2 learners’ use of standard German, dialect, and mixture varieties in interaction with a virtual standard German-speaking interlocutor and a virtual dialect-speaking interlocutor, at both the group and the individual level. In the inter-individual analysis, the findings showed that occupational complexity with things was related to participants’ use of dialect varieties, though the effect was comparatively weak. Since this effect was exclusively found in interaction with the standard German-speaking interlocutor, this may be an indication that, for L2 speakers whose occupations require more physical precision work, dialect use may be durable against macro-situational factors such as the variety of the interlocutor. Moreover, these findings are consistent with the notion that L1 Austro-Bavarian-speaking individuals in manual labor careers tend towards dialect varieties cross-contextually (Scheuringer 1990; Steinegger 1998; Vergeiner 2021) and also with previous variationist SLA work underscoring that L2 speakers with higher physical requirements may produce more informal dialectal variants (Prediger 2016). However, there were also critical cases of disparity. Our second goal was thus to perform a more ecological person-centered analysis that meets the criteria for an appropriately fine-grained investigation of individual idiosyncrasy (Geeslin et al. 2013; Kanwit 2019; Regan 2013; Wirtz and Pfenninger 2023b), which is an inherent part of SLA in adulthood (Pfenninger et al. 2023; Singleton and Pfenninger 2019).

The person-centered analysis offered complementary insights regarding the role of occupational complexity on participants’ observed interpersonal varietal behavior. Whereas we could identify from an individual-level perspective that participants with occupations comprising more manual handing and physical precision work evinced clearly dialectally colored varietal repertoires, we also found indications that individuals engaged in dialect and mixture varietal use regardless of the operationalized occupational complexity. For example, a teacher, a salesperson, a PhD candidate, and a medical student also evinced dialectally influenced varietal behavior, the rationales for which are likely complexly interwoven with psychosocial and sociological drivers (Wirtz and Pfenninger 2023a; Wirtz et al. 2024). Additionally, the person-centered analysis was beneficial in identifying several methodological shortcomings of the occupational complexity metric for sociolinguistic and SLA applications, particularly as concerns the measure’s cultural and contextual sensitivity.

As Eckert (1997) argues, work-related benchmarks and events are perhaps some of the most pervasive across adulthood, and are inextricably related to life experiences that shape the way an individual comes to draw on language (variation) as a communicative tool. However, it is notoriously difficult to capture the effects of occupation on language change and use (Forrest and Dodsworth 2016); this is exacerbated by the fact that occupations are not necessarily categorical, but rather house a dense bandwidth in sociological phenomena and complexity dimensions that are not subject to clear-cut bounds, categories, or thresholds. Thus, adhering to operational definitions of different complexity dimensions on continuous scales (Pfenninger and Festman 2021) rather than categorical ones can allow us to more faithfully capture the granularity in the effects of occupation, and moreover allow us to see which complexity aspects of occupation impact most notably on variable patterns. This point in particular is also of interest more broadly in sociolinguistics. As mentioned previously, socioeconomic status indicators (e.g., aggregated education level, family income, and occupational rank) are widely used as independent variables in sociolinguistics (and SLA), but these function only as proxies for occupation or occupational status and the complexity involved therein. Drawing on occupational complexity presents a way to gauge more directly the extent to which occupation-related variables predict both inter- and intra-individual patterns of sociolinguistic variation. In other words, more attempts need to be made to tease apart the differential effects of occupation-distinct variables and other socioeconomic status indicator predictors.

That said, the present investigation is subject to several drawbacks. Perhaps the biggest drawback of the present study is the bandwidth in occupations represented in this sample. Of the 36 participants, 15 were students, for whom, as previously mentioned, an occupational complexity measure was assigned based on the guidelines for career classification outlined in the DOT. For all intents and purposes, this measure still represents the complexity involved in their current “career”, but the large sample of students reduces the amount of variance present in the data. Future work should thus strive to include samples that are more heterogeneous in terms of occupations, in line with Long’s (2022: 430) call to include L2 learners who “reflect a broad diversity of linguistic, cultural, and social backgrounds”. What is more, the occupational complexity measure is still comparatively categorical. Equally, the occupational complexity measures that we used were developed for the US labor market (rather than the Austrian one) and last updated during the early 1990s. This leaves open questions regarding the cultural sensitivity of the measure: as the person-centered analysis showed (and the group-level analysis necessarily aggregated over), there may be occupations (e.g., teachers) in which previously accepted norms, such as communicatively and interpersonally more complex occupations being associated with standard language (e.g., only standard German may be spoken at school), no longer hold or are no longer modern, which likely necessitates a more multi-varietal repertoire on the part of the migrant. This is a notable drawback of the occupational complexity measure (specifically for its applications in sociolinguistics and SLA), especially because the complexity metrics with people and things are not context-specific and thus cannot countenance differences in, for example, interlocutor-related factors (e.g., interactional complexity with coworkers vs. students in the case of teachers) or physical precision-related factors (e.g., physical precision in working with machines vs. with instruments). In light of these drawbacks, it would appear to be an important next step for sociolinguistic and variationist SLA work to develop more fine-grained, sociometrically validated scales that capture the granularity of occupational complexity in terms of the communicative, manual, social, and psychological requirements of individuals’ primary occupations. Such scales have the potential to propel our knowledge forward concerning how we can “model the career and its possible interactions with other classic factors such as the social class and speaking style” (Riverin-Coutlée and Harrington 2022: 49) and, more generally, how occupational life-course trajectories impact on patterns of language variation.