Pre-service teachers’ ability to identify academic language features: the role of language-related opportunities to learn, and professional beliefs about linguistically responsive teaching

Abstract In order to reduce language-related disparities in educational attainment, teaching standards in many (Western) countries nowadays request teachers of all subjects and grades to focus on their students’ academic language development: Especially second language learners and socioeconomically disadvantaged students need continuous and systematic language support (linguistically responsive teaching) in order not to be left behind. To be able to act accordingly, (prospective) teachers need to acquire pedagogical language knowledge and develop positive beliefs about linguistically responsive teaching practices. Our cross-sectional survey among German pre-service teachers (n = 115) shows that – despite positive beliefs towards linguistically responsive teaching – hardly any of the participants had acquired a solid knowledge base in the area of educational linguistics: The majority could not identify a substantial amount of academic language features (known to be challenging especially for second language learners) in a content-specific explanatory text. This ability, however, is the basis for planning and carrying out linguistically responsive lessons. Although latent profile analyses show differences among the participants, many of them did not have opportunities to learn in the area of linguistic diversity during their studies. The results give rise to the question whether and how pedagogical language knowledge could become compulsory in teacher training. ZUSAMMENFASSUNG Um sprachbedingte Unterschiede im Bildungserfolg von Lernenden zu verringern, ist in den Bildungsstandards und Curricula vieler (westlicher) Länder mittlerweile festgeschrieben, dass Lehrkräfte aller Fächer und Klassenstufen die (bildungs-)sprachliche Entwicklung ihrer Schüler:innen im Unterricht unterstützen sollen: Insbesondere Deutsch als Zweitsprache Lernende und sozioökonomisch benachteiligte Schüler:innen benötigen eine durchgängige und systematische Unterstützung in Form von sprachförderlichem Fachunterricht, um ihre schulischen Erfolgschancen zu verbessern. Um dieser Aufgabe gerecht werden zu können, müssen (angehende) Lehrkräfte pädagogisches Sprachwissen erwerben und positive Überzeugungen zu einer sprachbewussten Unterrichtgestaltung entwickeln. Unsere querschnittlich angelegte Befragung unter Lehramtsstudierenden in Deutschland (n = 115) zeigt, dass trotz positiv beschaffener Überzeugungen zu sprachsensiblem Fachunterricht nur die wenigsten der Teilnehmenden über ein solides linguistisches Basiswissen verfügten: Die Mehrheit von ihnen war nicht in der Lage, sprachliche Merkmale, die sich insbesondere für Zweitsprachenlernende als verstehenserschwerend erwiesen haben, in einem fachspezifischen Erklärtext zu identifizieren. Diese Fähigkeit ist jedoch eine wichtige Voraussetzung dafür, Unterricht sprachsensibel zu gestalten. Latente Profilanalysen fördern zwar individuelle Unterschiede zwischen den Teilnehmenden zu Tage, es lässt sich aber festhalten, dass viele von ihnen während des Studiums kaum Lerngelegenheiten zum Umgang mit sprachlicher Diversität im Unterricht wahrgenommen haben. Die Ergebnisse werfen die Frage auf, ob und wie pädagogisches Sprachwissen ein verpflichtender Bestandteil der Lehrkräfteausbildung werden kann. PLAIN LANGUAGE SUMMARY In order to reduce language-related disparities in educational attainment, teaching standards in many Western countries nowadays request teachers of all subjects and grades to focus on their students’ language development: Especially second language learners and students growing up in socioeconomically disadvantaged families need continuous and systematic language support (linguistically responsive teaching) in order not to be left behind. To fulfil this challenging task, teachers do not only need to be willing to engage in such teaching practices, but have to develop awareness about the function and the particularities of academic language. In our online survey, 115 German pre-service teachers were asked to identify linguistic features that may potentially hinder high school students’ understanding of the content in a mathematical explanation. We also investigated the language-related learning opportunities the participants had during their university studies, and their attitudes towards linguistically responsive teaching. The participating pre-service teachers showed great awareness of linguistically responsive teaching being generally important and necessary. At the same time, their ability to identify academic language features was very limited. As knowledge about academic language is an important basis for planning and providing linguistically responsive lessons, it is questionable whether the future educators will be able to cater for the needs of linguistically diverse students. The rather worrisome result does not come as a surprise when looking at the pre-service teachers’ language-related learning opportunities: Many of the participants had not or only briefly encountered topics such as second language learning and linguistic diversity during the course of their university studies. The results give rise to the question whether and how pedagogical language knowledge can become a compulsory part of teacher education programs.


PLAIN LANGUAGE SUMMARY
In order to reduce language-related disparities in educational attainment, teaching standards in many Western countries nowadays request teachers of all subjects and grades to focus on their students' language development: Especially second language learners and students growing up in socioeconomically disadvantaged families need continuous and systematic language support (linguistically responsive teaching) in order not to be left behind.To fulfil this challenging task, teachers do not only need to be willing to engage in such teaching practices, but have to develop awareness about the function and the particularities of academic language.In our online survey, 115 German pre-service teachers were asked to identify linguistic features that may potentially hinder high school students' understanding of the content in a mathematical explanation.We also investigated the language-related learning opportunities the participants had during their university studies, and their attitudes towards linguistically responsive teaching.The participating pre-service teachers showed great awareness of linguistically responsive teaching being generally important and necessary.At the same time, their ability to identify academic language features was very limited.As knowledge about academic language is an important basis for planning and providing linguistically responsive lessons, it is questionable whether the future educators will be able to cater for the needs of linguistically diverse students.The rather worrisome result does not come as a surprise when looking at the pre-service teachers' language-related

Introduction
Empirical studies have repeatedly shown that skills in the language of instruction (especially academic language skills) are vital for content learning (e.g.Heppt et al., 2016;Townsend et al., 2012).The language demands of schooling are challenging for all students, particularly for second language learners and students growing up in socioeconomically disadvantaged families.As the (linguistic) diversification of the student population has increased in most Western countries in the last decades, teachers face the challenge of how to support these students in meeting the language, literature, and learning demands of educational standards.In Germany, educational curricula nowadays expect teachers of all subjects and grades to focus on their students' academic language development (Kultusministerkonferenz, 2017).At the same time, many mainstream classroom teachers have had little or no preparation for providing learners with the language-related assistance their students need.Therefore, the question arises of what teachers should know and be able to do in order to teach linguistically diverse learners effectively.
Educational research on teacher professionalism attributes an important role to teachers' knowledge and beliefs for their professional competence (e.g.Baumert & Kunter, 2006;Kaiser & König, 2019).However, language-related knowledge and beliefs are hardly mentioned in current models of teacher professionalism in Germany.A look beyond the country's borders shows that other frameworks for understanding teaching and learning (e.g.Bransford et al., 2007) recognize that knowledge about language is vital for the preparation of teachers in a changing world: In order to promote students' acquisition of language skills necessary for content learning, teachers need to develop what Galguera (2011) and Bunch (2013) have described as pedagocial language knowledge (PLK).Even though it is still being discussed what kind of language-related knowledge and skills PLK should comprise, there seems to be no doubt that (a) teachers need to be offered adequate opportunities to learn (OTLs) in their pre-service teacher preparation programs and in professional development in order to develop PLK and positive beliefs about linguistically responsive teaching (LRT) (e.g.Hammer & Berkel-Otto, 2019;Paetsch et al., 2019), and (b) foundational knowledge about linguistics and second language acquisition is an important base for linguistically responsive pedagogical practices.The latter includes developing awareness about differences between spoken and written language and the ability to identify linguistic features (academic language features) of disciplinary texts and tasks that may pose a challenge especially for second language learners and that have been described by several authors for different languages (e.g.Gogolin & Duarte, 2016;Schleppegrell, 2004;Zwiers, 2014).
Against the described backdrop, the aim of the present study is to shed light on the questions of (a) how well pre-service teachers in Germany are able to identify academic language features in a subject-specific explanatory text, (b) what kind of language-related OTLs they are offered at university, and (c) how their beliefs about LRT are shaped.By applying a person-centred approach, we identify different types of pre-service teachers based on their beliefs and language-related OTLs which are crucial for developing knowledge about academic language.The empirical insights provide practical implications about what prospective teachers need in order to be adequately prepared for dealing with linguistic diversity in the classroom.

The importance of (teaching) academic language
There is broad consensus that the development of strong language skills is vital not only for academic achievement, but also to be able to participate fully in a democratic society (Darling-Hammond & Bransford et al., 2007).Research has shown that it is necessary to differentiate between two differing kinds of language proficiency: Conversational language skills are required in social situations and can be classified as a rather colloquial, context embedded style of language.Contrarily, academic language (e.g.Cummins, 1991;Zwiers, 2014) or the language of schooling (Schleppegrell, 2004) is a formal, context-reduced, lexically dense, language register which serves the purpose 'to present information authoritatively, in a highly structured fashion' (Schleppegrell, 2001, p. 451). 1  Mastery of academic language is of importance for academic success due to several reasons: Academic language is used in textbooks, tasks, and classroom talk; it is the medium through which new content is passed on (Morek & Heller, 2012).Therefore, students have to be able to understand academic language (receptive skills).In addition, students are expected to use academic language when writing, giving oral presentations in class, and in exams (productive skills) in order to demonstrate their ability to participate in a discussion or present ideas in an adequate and educated manner (Morek & Heller, 2012).As cognitive development and language learning are closely interlinked, academic language also has an epistemic function (as a tool of thinking).Academic language acquisition is conjoined with the acquistion of subject-specific as well as subject-independent concepts and procedures: Scientifically complex and abstract concepts and processes cannot be represented completely by using everyday language, but need a kind of language matching their level of abstraction and complexity (Halliday, 1993).As part of the socio-cultural dimension of academic language acquisition, students have to learn which kind of language use is appropriate in different contexts to speak or write about different topics (e.g.Gebhard et al., 2011).
Most students do not learn academic language through mere exposure within a classroom setting; the acquisition of academic language is especially challenging for all children from families with home literacy practices different from school (e.g.second language learners, monolingual students growing up in low SES families, speakers of non-standard varieties and dialects).In order to help students master the dual challenge of learning the 'new language' and learning in it at the same time, they need to be provided with continuous and systematic support (Gibbons, 2002).The latter includes that mainstream teachers explicitly focus on the academic language development of their students and that combined language and content learning is provided across the curriculum.There are several didactic concepts and methods for the integration of language and content learning, for instance Sheltered Instruction (SIOP) (Short & Echevarria, 2015), Scaffolding (Gibbons, 2002), Linguistically Responsive Teaching (LRT) (Lucas et al., 2008), or Inclusive Academic Language Teaching (IALT) (EUCIM-TE Consortium, 2010).Even if there is still a lack of experimental and quasi-experimental studies examining the effectiveness of these teaching approaches (Becker-Mrotzek et al., 2021), empirical research supports the assumption that designing subject lessons in a way that language learning is promoted is a promising starting point for reducing disparities in academic achievement.

Teachers' professional competence (in the context of linguistic diversity) -the role of professional knowledge and beliefs
Although educational policies in most German federal states set a normative framework for practising LRT, empirical findings suggest that such teaching approaches have not been systematically implemented in German classrooms even though teachers generally display rather positive attitudes towards language-supportive teaching (e.g.Becker-Mrotzek et al., 2012;Brandt, 2021;Riebling, 2013).In order to explain their practice, teachers often refer to insufficient qualifications in the field of German as a second language and dealing with linguistic heterogeneity.
Educational research on teacher professionalism attributes an important role to teachers' knowledge and beliefs (in addition to other affective-motivational components) for their professional competence which affects instructional practices and, in turn, student outcomes (e.g.Baumert & Kunter, 2006;Kaiser & König, 2019;Kunter et al., 2011).Models to explain teachers' professional competence (e.g.Kaiser & König, 2019) therefore represent a suitable starting point for examining the complex interplay between teachers' professional knowledge, beliefs, and classroom action in the context of linguistic heterogeneity.
The taxonomy of teachers' professional knowledge proposed by Shulman (1986Shulman ( , 1987) has become widely accepted in educational research (e.g.Baumert & Kunter, 2006;König et al., 2014;Krauss et al. 2008;Lipowsky 2006): content knowledge (CK), pedagogical content knowledge (PCK), and pedagogical knowledge (PK) are considered to be core categories of teachers' professional knowledge (e.g.Baumert & Kunter, 2006;König et al., 2014).The tripartite division is reflected in the structure of German teacher training programmes, which are usually divided into subject-specific, subject-specific didactic, and pedagogical parts (Voss et al., 2015).Language-related knowledge, on the contrary, still only plays a subordinate role in the general discussion about teacher professionalism in Germany.Following Shulman's taxonomy, Galguera (2011) and Bunch (2013) have introduced the term pedagocial language knowledge (PLK) to describe the kind of language knowledge that teachers should develop in addition to CK, PCK, and PK.Other authors have suggested to use the terms educational linguistics (Fillmore & Snow, 2018) or disciplinary linguistic knowledge (DLK) (Turkan et al., 2014).Even though there is no uniform definition of what it is exactly that teachers need to know about language, existing approaches seem to agree on the fact that foundational knowledge about linguistics and second language acquisition and awareness about differences between oral and written language as well as the peculiarities and functions of academic language are an important basis for linguistically responsive pedagogical practices (Bunch, 2013;Lucas et al., 2008;Valdés et al., 2007).In their framework for linguistically responsive teacher education, Lucas and Villegas (2013) -among many other aspects -explicitly point to the fact that teachers need to develop 'skills for analysing the language of their academic disciplines and the linguistic demands of classroom activities […]' and 'the ability to examine language-related factors that influence teaching and learning, such as sentence structure and complexity in academic writing, types of cohesive devices used in different genres […], and specialized uses of vocabulary in specific disciplines' (p.106).Likewise, 'knowledge of language' and the 'description of language in linguistic terms' are described as being necessary for 'successful language-sensitive subject instruction' in the European core curriculum for inclusive academic language teaching (EUCIM-TE Consortium, 2010, pp. 20-22).Within the framework of scaffolding (Gibbons, 2002), the identification of the language demands of a given topic is a central element of lesson planning (macro scaffolding) which serves as the basis for implementing adequate language support strategies.Knowledge about grammatical structures and vocabulary as well as semiotic systems also forms an integral part of a structure model of teachers' German as a second language (GSL)-related competence (DaZKom model) which provides the theoretical basis for a standardized test instrument (DaZKom test = GSL competency test) measuring (pre-service) teachers' skills in dealing with GSL and multilingualism (Carlson et al., 2018).
In addition to the various dimensions of professional knowledge mentioned above, teachers' professional beliefs on teaching and learning have received special attention in teachingand learning-related research, because it is assumed that they are also closely related to classroom practice and can thus influence the quality of teaching and the learning success of students (e.g.Buehl & Beck, 2014;Pajares, 1992).Beliefs are (usually) non-scientific ideas, views, and opinions that are subjectively believed to be true, even if they are contradicted by reasoned arguments or epistemic knowledge (e.g.Oser & Blömeke, 2012;Richardson, 1996).They are usually not neutral, but involve an evaluation of things, people, and situations (nespor, 1987;Seiz et al., 2017).Teachers' professional beliefs can be defined as notions and assumptions about school-and instruction-related phenomena and processes (Kunter & Pohlmann, 2009).Teacher education frameworks with a focus on LRT and inclusive academic language teaching in the United States (Lucas & Villegas, 2013) and Europe (EUCIM-TE Consortium, 2010) also emphasize that teacher beliefs are fundamental for successfully dealing with linguistic diversity in the mainstream classroom.

Language-related knowledge of teachers
Even though empirical findings on the role of teachers' knowledge and its influence on classroom practices in the context of linguistic diversity are still scarce (Kalinowski et al., 2020), a growing body of studies in Germany has set their focus on (prospective) teachers' language-related knowledge: During the standard-setting procedure of the abovementioned DaZKom test (N = 1,383), about 91 percent of the participating pre-service teachers of different subjects did not achieve a minimum standard of GSL competence (Gültekin-Karakoç, 2018).In a similar study, almost 90 percent of pre-service STEM teachers (N = 298) did not have sufficient GSL-related knowledge (Schroedler & Lengyel, 2018).
Hardly any empirical research specifically focusses on (prospective) teachers' knowledge about academic language and its features.A qualitative think-aloud study by Petersen and Peuschel (2020) among pre-service teachers (N = 16) without special training in the area of GSL indicates that the participants mainly notice features of linguistic complexity on the word-level, particularly subject-specific vocabulary.A quantitative study by Jost et al. (2017) among in-service teachers (N = 340) concludes that most educators are generally aware of language posing a barrier to their students' understanding of the subject content.A smaller in-depth study (n = 5) reveals that maths teachers also identify subject-specific vocabulary as the main linguistic challenge in mathematical tasks whereas language support coaches also show awareness about other lexical and grammatical linguistic phenomena (Jost et al., 2017).A third study by Wallner (2020) shows that subject-specific vocabulary is the most prominent academic language feature identified by in-service teachers completing a GSLqualification as part of their professional development (N = 95).
To sum up, existing empirical research points to the fact that many (prospective) educators in Germany do not have sufficient language-related knowledge that would be necessary for providing effective LRT.The recognition of language barriers in texts and tasks generally seems to be limited to surface-level linguistic information.

Opportunities to learn and professional teacher beliefs (in the context of linguistic diversity)
The university education of prospective teachers is a central place to lay the foundation for the acquisition of PLK by offering appropriate OTLs.However, whether and to what extent topics such as language education, dealing with linguistic heterogeneity, and GSL are part of the teaching canon in the first phase of teacher training varies between the different German federal states and depends on the studied subjects and the intended type of teacher qualification (e.g.high-school vs. primary school) (Baumann, 2017). 2 In surveys, teachers report not to feel adequately prepared to work with linguistically diverse students by the teacher training they completed (e.g.Becker-Mrotzek et al., 2012;Pettit, 2011).A study by Ehmke and Lemmrich (2018) among pre-service teachers (N = 496) at 12 German universities comes to the conclusion that GSL-related content only plays a minor role in teacher training: More than half of the participants have not or at most in one session been introduced to topics such as academic language or scaffolding during the course of their studies.Even fewer of the prospective educators have carried out specific GSL-related activities such as the analysis of texts concerning their linguistic complexity or the planning of linguistically responsive lessons.A positive effect of learning opportunities (in the form of PD) on in-service teachers' ability to provide content and language integrated learning has been revealed in a meta-analysis by Kalinowski et al. (2020).Similarly, Babinski et al. (2018) showed in a randomized controlled trial that PD with a focus on high-impact instructional strategies positively influences students' language and literacy skills.
Empirical evidence further suggests that teachers in Germany generally hold very positive beliefs about LRT when the support of the language of instruction (e.g.German in Germany) is concerned (e.g.Fischer et al., 2018;Schroedler & Fischer, 2020): Surveys among (pre-service) teachers reveal that (prospective) educators largely agree that all lessons should be language lessons (Tajmel, 2010, N = 125) and that language support is generally useful and necessary (Becker-Mrotzek et al., 2012, N = 512).Including multilingual practices in the mainstream classroom, on the contrary, is seen rather critically by many (e.g.Brandt, 2021;Putjata, 2018;Schlickum, 2013).Strong monolingual beliefs have been shown to be present among teachers from many other countries around the globe (e.g.French, 2016;Pulinx et al., 2016;Young, 2014).Based on self-reports of in-service social science teachers (N = 198), a study by Brandt (2021) provides empirical evidence for the assumption that both teachers' professional beliefs about LRT and the participation in GSL qualifications have an impact on their classroom practice.

Relationship between teachers' language-related knowledge, opportunities to learn and professional beliefs
Empirical research on the relationship between (pre-service) teachers' knowledge, their beliefs, and OTLs in the context of linguistic diversity indicates a positive relationship between the amount of OTLs and teachers' beliefs about LRT, i.e. the more OTLs in the area of GSL and dealing with linguistic heterogeneity, the more positive their views about LRT (and vice versa) (e.g.Gándara et al., 2005;Lee & Oxelson, 2006;Mantero & McVicker, 2006;Schroedler et al., 2022).Also, beliefs about LRT are positively associated with GSL competence (e.g.Hammer et al., 2016;Schroedler & Fischer, 2020), and the amount of OTLs in the area of GSL (inside as well as outside university) predicts the GSL-related competence of preservice teachers (e.g.Ehmke & Lemmrich, 2018;Paetsch et al., 2019).

Research questions
There is growing evidence for the importance of LRT and related OTLs in teacher education.Less is known, however, about pre-service teachers' ability to identify linguistic features that might be challenging for learners and should thus be considered when planning and conducting lessons.Moreover, the relationship between pre-service' teachers ability to detect linguistic challenges, their OTLs, and their beliefs about LRT requires further research.Therefore, the current study aims to answer four research questions: 1. To what extent are pre-service teachers able to identify academic language features in a mathematical explanation?2. What kind of OTLs in the area of GSL and dealing with linguistically diverse learners in class do pre-service teachers encounter during the course of their studies?3. What are pre-service teachers' beliefs about linguistically responsive maths teaching?4. Are there different types of pre-service teachers regarding their ability to identify academic language features, their beliefs about linguistically responsive teaching, and the amount of language-related OTLs they had during their studies?

Method
In order to address the research questions, we use data from a cross-sectional online-survey carried out among students being enrolled in a teacher-education program (primary and intermediate secondary school-bound) 3 at a German university without compulsory study elements on language education or linguistic diversity in December 2020. 4The study was conducted during two online-lectures and took approximately 40 minutes to complete (there was no time limit).However, participation was voluntary and anonymous.Out of 135 students, who initially participated in the study, 17 dropped out before the end of the survey.The answers of another three participants were excluded from the study as implausible answering patterns were detected.The following sample description is on the 115 cases used in our analyses.

Sample and instruments
Most of the pre-service teachers participating in the study (94.9%) were female.The majority of students (71.8%) were between 21 and 25 years old.12.8% grew up with a language (or dialect) other than standard German; 8.5% had a so-called migrant background. 5Most of the students (59.0%) were completing their master's degree, and 41.0% were still in the bachelor phase.About three quarters (76.5%) of the students wanted to become primary school teachers while the remaining 24.5% were planning to teach at intermediate secondary schools (Haupt-/Realschulen). Roughly a third of the students (38.2%) studied mathematics and 58.1% German as one of their subjects.About every fifth participant (19.7%) had participated in a GSL-qualification programme before or was enrolled in one at the time of the survey.
The survey consisted of six parts: First, the students were asked to read a mathematical explanation and to rate its overall linguistic complexity (also see Strohmaier et al., 2023).Second, they were exposed to five different parts of the explanatory text (consisting of three to five sentences each) and asked to identify all of the linguistic features that may potentially pose an obstacle to understanding the explanation for ninth-graders (open-ended questions) (see Appendix 1).In part three, the pre-service teachers' own understanding of the mathematical explanation was tested with nine multiple choice items (α = .71).Part four focussed on OTLs in the area of GSL and linguistic diversity at university.Part five measured the pre-service teachers' beliefs about LRT.In the last part, the participants answered questions on their personal and professional background.
The text blocks the students were asked to analyse in part two of the assessment were linguistically manipulated and included a total of 44 academic language features which can be assigned to five sub-dimensions: (a) syntactic features (e.g.subordination and complex sentence structure), (b) morphosyntactic features (e.g.genitive constructions), (c) impersonal expressions, (d) referential ambiguity, and (e) lexical features (subject specific as well as general academic vocabulary). 6Based on a coding manual, the students' open-ended responses were manually coded by two researchers independently.The interrater reliability was excellent κ = .90(Cohen, 1960).Remaining intercoder disagreements were resolved by a third coder.POMP (Percent of the Maximum Possible) scores of the correctly identified academic language features were calculated.In POMP, the score assigned to each individual is a percentage, reflecting the individual's position on the scale as a Percent of the Maximum Possible score achievable on the scale (Cohen et al., 1999).
The students' OTLs in the area of GSL and dealing with linguistically diverse students were assessed with two scales adapted from Ehmke and Lemmrich (2018): The first scale focussed on 'thematic OTLs' , for example, 'During the course of your university studies, how often were topics such as linguistic diversity addressed?' .The scale consisted of nine items (α = .94).The five-point response scale reached from 1 = 'never' to 5 = 'in several complete courses' .A second scale, 'GSL-related activities' , included items that specifically asked about GSL-related activities practiced in university courses, e.g.'During the course of your university studies, how often did you analyse texts concerning their linguistic complexity?' .The answering options also reached from 1 = 'never' to 5 = 'in several complete courses' .The scale consisted of five items (α = .74).The scale reflecting prospective teachers' beliefs on how to deal with linguistic diversity in the maths classroom was adapted from the DaZKom belief survey (Fischer et al., 2018;Lahmann & Fischer, 2020).It consisted of 10 items measuring pre-service teachers' beliefs about linguistic responsiveness and responsibility for language facilitation, for example, 'Academic language skills are an important basis for content learning in the mathematics classroom' .Response categories were coded as follows: 1 = 'don't agree at all' to 4 = 'fully agree' .negatively worded items were recoded so that a higher mean score on the scale equates to a more positive view towards LRT (α = .69).
The unidimensionality of all scales was tested with exploratory factor analyses (principal-component factors) using Stata 16: The Kaiser-Meyer-Olkin measure verified the sampling adequacy for the analyses, and all KMO values for individual items were greater than .63.The inspection of scree plots (Cattell, 1966), the Kaiser criterion (Kaiser, 1960), and Horn's parallel analyses (Horn, 1965) suggested that one-factor solutions best fit the data.The factors explained 28% (beliefs), 52% (GSL-related activities), and 70% (thematic OLS) of the total item variance.

To what extent are pre-service teachers able to identify academic language features in a mathematical explanation?
Figure 1 shows the percentages of correctly identified academic language features in the five text sections.13.9% of the respondents did not identify any of the linguistic features as potentially posing an obstacle to understanding the text; more than a third of the students (38.3%) identified between one and 10%.Fewer than ten percent of the participants (6.1%) were able to detect more than 30% of the linguistic difficulties that the texts contained.
Figure 2 shows the percentage of correctly identified academic language features separately for the different dimensions of linguistic complexity described in the theoretical background section. 7The percentages of correctly identified academic language features in the categories 'syntax' (21%), 'content-specific academic language vocabulary (ALV)' (19%) and 'general academic language vocabulary (ALV)' (15%) appear to be much higher than those of the categories 'morphosyntax' (3%), 'referential complexity' (3%), and 'passive constructions/impersonal expressions' (1%).
A Friedman AnOVA confirms that the overall difference between the categories is statistically significant (χ 2 (5) = 234.54,p < .001).Follow-up pairwise comparisons (Wilcoxon tests with Bonferroni corrections) show that the differences between the first three dimensions (syntax, content-specific, and academic language vocabulary) and the last three dimensions (morphosyntax, referential complexity, and passive/impersonal expressions) are statistically significant (p < .01)while the differences within the first and within second set of categories are not (p ≥ .05).

What kind of OTLs in the area of GSL and dealing with linguistic diversity in class did pre-service teachers encounter during the course of their studies?
Figure 3 shows which and how many thematic OTLs in the area of GSL and dealing with linguistic diversity the participating students had during the course of their university studies.A comparison of the students' answers on the item level shows that the amount of OTLs students had during the course of their studies varies immensely between the different topics: While a large proportion of the participating students (over 40%) reports to have 'never' or at most 'in one session' been confronted with topics such as 'academic language acquisition' or 'language support' , about half of the participants report that 'second language acquisition phenomena' (49.6%) and 'German grammar' (56%) have been addressed in a complete course or several complete courses.About two thirds of the participating pre-service teachers (66.1%) had dealt with 'linguistic subdomains' in one or several complete courses.A mean scale score of 3.06 (SD = 1.14, min.= 1, max.= 5) suggests that -on average -students had dealt with GSL-related topics 'in several sessions' during the course of their studies.However, a standard deviation of 1.14 and a scale score range from 1 to 5 indicates that the students' experiences are rather heterogeneous.
In addition to the described thematic OTLs, students reported how often they had the opportunity to engage in specific GSL-related activities during the course of their study.The mean score of the scale is 2.1 (SD = 0.79) which means that, on average, the participating students had carried out GSL-related activities in 'one session' during the course of their university studies.It can therefore be concluded that the amount of specific GSL-related activities students had in the first phase of their teacher training is even lower than their experience with thematic OTLs in the area of GSL and dealing with linguistically diverse students.Again, the range of the scale scores (min.= 1, max.= 5) indicates that the number of GSL-related activities students had encountered differs greatly within the sample.
As shown in Figure 4, the experiences also vary immensely between the different GSLrelated activities: Almost 90% of the students state to have 'never' or at most 'in one seminar session' created individual language support plans, almost two thirds (64.4%) had 'never' or at most in 'one seminar session' planned linguistically responsive lessons, and more than half of the pre-service teachers (53.9%) had 'never' or at most in 'one seminar session' analysed texts concerning their linguistic complexity.More than half of the participants  (54.8%), on the contrary, had experience with diagnosing different stages of language acquisition.

What are pre-service teachers' beliefs about linguistically responsive maths teaching?
As can be seen in Figure 5, the participating prospective teachers hold very positive beliefs about LRT.The level of agreement is particularly high for the positively-worded items: Almost all of the participants (97.4%) (rather) agree that academic language skills are an important basis for content learning in the mathematics classroom and that teachers should take the linguistic competencies of their students into account when selecting tasks.
Slightly lower agreement levels are reached for the negatively-worded items: Almost a third (28.7%) of the participants is of the opinion that mathematical language can be acquired independently of everyday and academic language, and a little over one fourth (26.5%) (rather) agrees that mathematical problems can be solved without academic language skills.The distribution of the scale score (M = 3.4,SD = 0.33, min.= 2.4, max.= 4) is skewed to the left which reflects that the students' beliefs about LRT are generally very positive.

Are there different types of pre-service teachers regarding their ability to identify academic language features, the amount of OTLs they had during their studies, and their beliefs about LRT?
In order to answer Research Question 4, latent profile analyses were carried out with Stata 16.An initial run of two to four clusters was analysed, with the four indicator variables of interest: (a) the percentage of correctly identified linguistic features (POMP), (b) thematic OTLs, (c) GSL-related activities, and (d) the students' beliefs about LRT.As GSL-related activities did not contribute to the identification of profile membership, they were removed and the analyses rerun.In order to determine which of the three possible models fits the data best, Akaike's information criterion (AIC), Schwarz's Bayesian information criterion (BIC), and log-likelihood values were inspected (with lower absolute values indicating better fit).The three-profile solution has the smallest AIC (152.05),BIC (179.50), and log-likelihood (−66.03)values when compared to the two-profile (AIC = 245.22;BIC = 272.67;LL = −112.61)and four-profile solution (AIC = 214.33;BIC = 263.74;LL = −89.17).Based on these criteria, the three-profile solution can be considered the best model.The marginal means of the observed variables of the three-profile solution are shown in Table 1.
In order to better illustrate the differences in the results of the three profiles, the three variables of interest (POMP score, thematic OTLs, and beliefs) have been z-standardized.The standardized mean scores are displayed in Figure 6.The three profiles can be described as follows: Profile 1: Surprise -High performance and very positive beliefs despite a medium amount of learning opportunities Profile 1 represents only 9% of the participating students.They found (on average) one third of the linguistic features that may possibly hinder students from fully understanding the given explanation.With an average score of 3.6 (with a theoretical maximum of 4; SD = 0.4), this profile is characterised by extremely positive beliefs towards LRT.Students in this group stated to have taken language-related OTLs in at least several seminar sessions during the first phase of teacher education.About half of the students in this profile (54.6%) studied German as a subject.
Profile 2: Against all odds -Low performers with many learning opportunities With almost half of the participating students (48%), Profile 2 is the largest group within our sample.Students in Profile 2 were able to identify one out of 10 features of linguistic complexity in the given mathematical explanation.At the same time, students belonging to this group reported the highest number of language-related OTLs of all profiles: They had been exposed to the topic in at least one complete seminar.With an average score of 3.5 (SD = 0.38), the beliefs about LRT in this profile are only slightly more negative than those in Profile 1.Most of the students in this profile (91.1%) were enrolled in German as a teaching subject.
Profile 3: As can be expected -A lack of learning opportunities and low performance With 43%, Profile 3 forms the second largest group.Students in this profile were hardly able to find any of the features of linguistic complexity in the mathematical explanation.On average, they reported to have had language-related OTLs in only one seminar session during the first phase of their teacher education.Even though the beliefs about how to deal with linguistic diversity in the mainstream classroom in this profile were also positive (M = 3.2, SD = 0.36), they were -on average -lower than those of students in Profile 1 and Profile 2. Only one fifth (20.83%) of students in this profile had German as a teaching subject.

Discussion
The present contribution provides evidence on how familiar pre-service teachers at a German university without compulsory study elements on language-supportive teaching are with features of academic language: The majority of the participants was only able to identify a very small number of academic language features in a given explanatory text.The result is in line with previous research that has attested low levels of language-related knowledge in pre-service teachers at different German universities.Differing from previous studies, our study focussed on a specific sub-skill of PLK -the identification of academic language features.The participants could identify linguistic obstacles mostly (if at all) on the word and syntax level while features on all other levels were hardly recognized.Our research therefore confirms the observation of previous qualitative studies that (pre-service) teachers' awareness tends to be restricted to obvious features on the linguistic surface while phenomena on deeper levels remain largely uncovered.That students were relatively good at correctly identifying general academic vocabulary (≈ 15%) is somehow surprising as it is often discussed as being more difficult to recognize than subject-specific academic vocabulary.However, a closer examination of the individual items within the category 'general academic vocabulary' shows that two out of the 10 items have been recognized much more frequently (about 50%) than the other words (≤ 10%).The fact that both words are foreign words derived from Latin ('exemplarisch ' [exemplary] and 'kalkulieren' [calculate]) may explain why they were detected more often by the participants than the other words in this category.Considering that the identification of the linguistic demands of a given topic and corresponding classroom activities is a central element of language-sensitive lesson planning, the overall results of our study cast doubt on whether the participating pre-service teachers' had acquired the knowledge necessary to provide linguistically responsive teaching.
Our descriptive analyses reveal that many of the prospective educators either were not offered or did not make use of OTLs related to GSL and dealing with linguistic diversity in class during the initial phase of their teacher training.Especially language-related activities, i.e. specific activities related to dealing with linguistically diverse students, had not been encountered by a large proportion of the participants during the course of their studies.The rather limited knowledge about academic language features and the low-usage of language related OTLs cannot be explained by the students' beliefs about how to deal with linguistic diversity in the mainstream classroom: In line with the results from previous research, the measured beliefs towards LRT were extremely positive, which implies that the participating pre-service teachers are generally aware about its importance.This finding is encouraging as teachers' professional beliefs have been shown to influence instructional practices and to affect student achievement.However, as our assessment of the pre-service teachers' beliefs relied on self-reports, we cannot rule out the possibility that the answers were influenced by social desirability.Hence, it is possible that the explicitly expressed beliefs differ from the pre-service teachers' actual (implicit) beliefs which usually occur outside of conscious awareness, but are known to play a decisive role in a teacher's professional development.
By choosing a student-centred approach and carrying out latent profile analyses, we could show that the participating pre-service teachers can by no means be characterized as a homogenous group.The three identified profiles differed immensely, especially concerning the students' ability to identify academic language features and the amount of language-related OTLs they took during the course of their university studies.While one of the observed patterns was quite in line with theoretical expectations -a lack of language-related OTLs going hand in hand with a low ability to identify academic language features -the combinations in the two other profiles were rather unexpected: Both, a (relatively) high ability to identify academic language features despite a small reported amount of OTLs as well as low test results in spite of a rather high number of OTLs contradict the theoretical (linear) assumption of 'the more of the one, the higher the other' and therefore require further examination.The rather paradox result that a small proportion of students was able to identify a high percentage of academic language features despite having taken relatively few language-related OTLs at university gives rise to the question whether more and/or other OTLs, especially outside of university, should be included in further research on the topic.It seems advisable to gather information about pre-service teachers' practical experiences in dealing with linguistic diversity, e.g. during teaching placements and tutoring programs, in order to gain a more comprehensive picture of the importance of learning opportunities.At the same time, the observation that a large group of students reported to have taken many language-related OTLs and still scored low in our assessment leads to the question of whether the items that were chosen to measure language-related OTLs, were too unspecific: Items such as 'dealing with linguistic subdomains' , or 'German grammar' -which are both mandatory when studying German as a teaching subject -do not necessarily fit to the test which focusses on a very specific (but important) facet of PLK, i.e. the identification of academic language features.

Limitations/further research
The results of our study help to gain a better understanding of what prospective teachers know about academic language, GSL-related learning opportunities they had during their university studies, and the beliefs they hold about LRT.nonetheless, the results must be interpreted with caution, and a number of limitations should be borne in mind.
Some limitations concern our sample: The study was carried out at one out of many institutions of teacher education in Germany.The number of participants was relatively small and did not include all school types.This is important because grammar-school teachers tend to hold more negative beliefs about LRT (e.g.Brandt, 2021), and pre-service teachers planning to teach at this type of school were not included in the sample.The results of our study are therefore not representative for all prospective teachers in Germany.Further studies should include and compare students from different universities with and without mandatory classes in the area of GSL/LRT and represent pre-service teachers of all school types.It would also be desirable to conduct similar studies in other countries in order to identify training conditions that foster (pre-service) teachers' professional competences in dealing with linguistic diversity.In addition, it should be noted that the participation in the study was completely voluntary and that there were no incentives.As a result, quite a few students dropped out of the assessment after a couple of minutes.We therefore assume that students who finished the survey had a genuine interest in the topic which could have positively influenced the results of our study.
Further limitations concern the measurement: In order to measure the pre-service teachers' knowledge about academic language features, we used a linguistically manipulated mathematical explanation.From a methodological perspective, the fact that some of the linguistic features were not identified by any (or very few) pre-service teachers indicates that these items were too difficult (low item discrimination).What may have contributed to the small amount of academic language features being identified is the fact that the students were not informed how many potential linguistic obstacles could be identified within each paragraph.This might have made it difficult to decide when to stop and move on to the next task while knowing how many features one should identify could have helped to persist and accomplish the task.In order to evaluate the appropriateness of the test and to better understand the results of the present contribution, the survey should be repeated with experts in the field of GSL/ linguistics and practising maths teachers.

Implications for practice
Despite being already in the master phase of their studies, most of the pre-service teachers in our study showed a (rather) limited ability to identify academic language features in a content-specific explanatory text.As this kind of language-related knowledge serves as a basis for being able to plan and provide effective linguistically responsive lessons, it is questionable whether these future teachers would be able to support students in the process of developing academic language skills and to teach in a linguistically responsive way.At the same time, it is quite obvious from our analyses that the participants had not been provided with the OTLs necessary to sharpen their awareness of academic language during the course of their studies.The latter also holds true for pre-service teachers being enrolled in German as a teaching subject.A practical implication of our study for (initial) teacher education therefore is that academic language, its characteristics and functions in different content areas should become mandatory topics for pre-service teachers of all subjects and all school forms in initial teacher education.In addition to basic linguistic knowledge, the importance of language for the respective subject and concept-building should be conveyed and the linguistic development of learners should be considered as part of the general education process.Both building up a substantial amount of knowledge about the relevance of language for content learning as well as the reflection of beliefs towards linguistically diverse students and LRT seem to be necessary to enable future teachers to facilitate combined content and language learning and prepare them for dealing more competently with some of the teaching and learning-related challenges in our fast-changing world.

Notes
4. The university does not offer training for teaching in academic secondary schools (Gymnasium).5. Either the student him-/herself and/or one or both parents born outside of Germany.6. Apart from literature presented in theoretical background section, the linguistic manipulation of the text was based on empirical findings about academic language features of German that have proven to potentially impede students' understanding of content (e.g.Haag et al., 2013;Heppt et al., 2014).7. Content specific and general academic language vocabulary (lexical features) are reported separately.

Figure 1 .
Figure 1.Percentage of correctly identified academic language features by percentage of students (n = 115).

Figure 2 .
Figure 2. Percentage of correctly identified academic language features by dimensions of linguistic complexity (with std.errors) (n = 115).

Figure 6 .
Figure 6.Three differing profiles of pre-service teachers (z-standardized mean scores with std.errors).

Table 1 .
Marginal means of the observed variables conditional on being in Profiles 1, 2 and 3 (n = 115).