One approach to integrating corpora into language teaching is to enable learners to conduct exploratory, autonomous research (cf. Flinz 2021: 4). This method allows students to engage with corpora independently, positioning them as researchers in their own right (cf. Johns 1988: 14) or as linguistic “Sherlock Holmes” (Johns 1997: 101). This was precisely the objective of the present project: to empower students to assume the role of linguists and conduct small-group linguistic research. In alignment with the course content, the project focused on the temporal development and semantics of COVID-related neologisms.

The students first selected relevant neologisms, including both established terms that had been incorporated into the online version of the Duden1 (and thus, to some extent, lost their neologism status) and unestablished ones. They then conducted both qualitative and quantitative analyses of these terms, examining their frequency since 2020 as well as their most common co-occurrences. The research aimed to address two central questions:

To answer these questions, the students interpreted their data, integrating hypotheses that the frequency of COVID-related neologisms would have declined between 2020 and 2023 (the year of the project) and that established neologisms would exhibit higher frequencies than unestablished ones.

For this project, the corpus was used directly in class within student groups, serving both as the foundation for their research and as a source of authentic language data, thereby exposing learners to natural language usage (cf. Lemnitzer / Zinsmeister 2015: 189). After collecting and analysing their data, the students presented their findings and wrote a critical review of the results.

In corpus-based language teaching, three primary approaches can be distinguished along the axes of empiricism/rationalism and deduction/induction (cf. Flinz 2021: 28): corpus-based quantitative, corpus-based quantitative and qualitative, and corpus-supported (cf. Lemnitzer / Zinsmeister 2015: 34). Given that this project combined frequency-based analysis with co-occurrence analysis and subsequent data interpretation, it followed a corpus-based quantitative and qualitative approach – widely regarded as the most suitable method for foreign language teaching (cf. Lemnitzer / Zinsmeister 2015: 37). This research-oriented methodology enabled learners to engage with language at a meta-level while providing an additional channel for linguistic input, complementing the examination of authentic texts (cf. Wallner 2013).

Aligned with data-driven learning (DDL; cf. Section 1), the project sought to enhance a range of transferable skills relevant to employability. According to the International Labour Organisation (ILO), individuals are most employable when they possess broad-based education and training, alongside essential portable skills such as teamwork, problem-solving, information and communications technology (ICT) proficiency, and strong communication and language skills (cf. Brewer 2013: iii). Other key employability skills include planning and organisation, self-management, autonomous working, and critical thinking (cf. ibid.: 1). Accordingly, the project was designed to develop all these competencies, with a particular focus on enhancing students’ metalinguistic awareness and practical German language skills through engagement with authentic German text data and an exploration of contemporary linguistic trends.

By positioning students as independent linguistic researchers, the project granted them considerable autonomy. While they received some guidance – such as pre-formulated research questions and references to useful resources – they were ultimately responsible for managing their time effectively, planning and organising their research, and ensuring timely completion. This autonomy was also intended to enhance learner motivation, as autonomy is recognised as one of the so-called ten commandments for fostering motivation in language learning (cf. Dörnyei / Csizér 1998).

Working in small groups required students to communicate and collaborate effectively to achieve successful outcomes. Thus, the project not only aimed to improve teamwork and communication skills but also sought to promote social learning. Additionally, by analysing and critically reviewing their findings, students were expected to further refine their critical thinking skills.

To present their data, students needed to familiarise themselves with Excel and PowerPoint-tools with which many had little to no prior experience, particularly in the context of creating figures and visualising data. Thus, beyond enhancing digital competencies, the project required students to engage with a new technological tool – a corpus – which was unfamiliar to all participants. This process helped develop their technological proficiency and independent learning skills. Furthermore, the presentation component of the project aimed to improve students’ public speaking abilities. Confronted with multiple new tools and concepts throughout the project, students also strengthened their problem-solving skills by navigating unfamiliar software and analytical methods.

Another crucial employability skill fostered by the project was linguistic proficiency. In addition to working with authentic German text data, students encountered new vocabulary and lexical fields while searching for relevant neologisms. During the frequency analysis, they had to determine whether corpus search results genuinely reflected neologism usage. For example, the word Sicherheitsabstand (‘safety distance’) existed before the COVID-19 pandemic – commonly referring to the distance between vehicles – and thus had to be excluded when it did not appear in a pandemic-related context. This process required students to critically analyse corpus data, differentiating between various contextual meanings of the same word.

Similarly, conducting a co-occurrence analysis further enhanced students’ contextual understanding of the selected neologisms. Through these tasks, they developed a more nuanced awareness of the semantic diversity of certain German words. Finally, by engaging with contemporary linguistic phenomena, students also gained insight into recent developments in the German language.

A wide range of corpora is available for German. Some of the largest and most well-known for written German include the Deutsches Referenzkorpus (DeReKo, ‘German Reference Corpus’)2 and the corpus from the Digitale Wörterbuch der deutschen Sprache (DWDS, ‘Digital Dictionary of the German Language’)3. For spoken German, the Datenbank für Gesprochenes Deutsch (DGD, ‘Database for Spoken German’)4 provides a comprehensive collection of data (for a more detailed overview of German corpora, see Flinz 2021: 9-28).

While the use of customised corpora in German as a Foreign Language (GFL) didactics offers advantages – most notably, allowing learners to work with smaller datasets, thereby reducing the risk of being overwhelmed by an excessive volume of data and failing to achieve meaningful results (cf. Flinz 2021: 12) – it was decided, given the linguistic focus of the course (cf. Section 2.3), not to create a customised corpus for this project. Instead, the Deutsches Referenzkorpus (DeReKo) was selected. As the world’s largest collection of German-language corpora, DeReKo comprises approximately 57.6 billion words (as of 9 January 2024; cf. IDS 2024). It encompasses a wide variety of texts, including fiction, academic and popular science works, newspaper articles, and other materials chosen based on their scope, quality, and topical relevance, with the majority originating from the present and recent past (cf. Flinz 2021: 10). The corpus is accessed via COSMAS II (Corpus Search, Management and Analysis System), an application used for searching and analysing the dataset (cf. COSMAS II). DeReKo is continuously expanded, growing by approximately 3.1 million words per year (cf. Kupietz et al. 2018: 4353). Given its scale and complexity, it is primarily designed for linguistic research rather than didactic purposes (cf. Flinz 2021: 11). However, this does not imply that it cannot or should not be used in language teaching (see, for example, Nolting / Radke 2019). Nevertheless, a certain amount of time is required to familiarise students with the corpus to ensure its effective use in educational contexts (cf. Steyer 2008: 189).

Although the introduction to the corpus is time-intensive, DeReKo was deemed appropriate for this project for several reasons. Firstly, the ability to navigate a new and complex tool is a crucial skill in the contemporary workplace (cf. Section 2.1). Thus, rather than perceiving the corpus’s complexity and unfamiliarity as obstacles, they were viewed as opportunities for students to develop transferable skills. Secondly, DeReKo provides an extensive and up-to-date dataset, which is essential for researching neologisms. Owing to its broad range of texts, students could investigate neologisms across different genres and registers. Additionally, the corpus allows users to filter results by decade, year, text source, topic, and other parameters – an essential feature for the temporal frequency analysis the students were required to undertake. Finally, DeReKo supports both frequency and co-occurrence analysis, which were central components of the project. Therefore, despite the time investment required for corpus training, DeReKo was selected as the most suitable resource.

The project was conducted over two consecutive academic years (2022/23 and 2023/24) as part of the Lingua Tedesca II course at the Università degli Studi di Urbino Carlo Bo in Italy. Second-year German students in Urbino follow a language course and a linguistics course for their German language component. They are required to pass four different exams annually. Related to the language course they have to pass a written exam (which consists of either a grammar test combined with writing and reading comprehension or an essay, depending on their proficiency level), a listening comprehension test, and an oral examination. For the linguistics course they need to pass a German linguistics exam. Lingua Tedesca II constitutes the linguistics component for all second-year German students. As such, its primary aim is to enhance students’ metalinguistic awareness rather than their practical language skills. This focus is also reflected in the course’s instructional language: while authentic German materials (e.g., videos and PowerPoint slides) are incorporated, the course is predominantly taught in Italian. The rationale for this approach lies in the diverse proficiency levels among students, who may range from A2 to C1 level upon entering their second year, depending on the level they started with in the first year. Consequently, while some second-year students operate at an A2 level, others may already have reached C1, making exclusive instruction in German impractical.

Attendance is not mandatory for the course, though passing the exam is a prerequisite for progression. Students have the option to participate in the project or to take an oral examination at the end of the term. Despite the non-compulsory nature of attendance, the majority of students opted to frequent the course regularly. In 2022/23, 23 students participated in the project, while in 2023/24, 34 students participated. The course takes place in either the first or second term and consists of 30 teaching hours, each lasting 45 minutes, in accordance with the academic quarter system.

The course content is determined by the lecturer. In the two years covered by this study, the course focused on neologisms and was structured in two parts: a theoretical introduction and a practical component centred around the project. The theoretical section was divided into three units: an introduction to neologisms (including definitions, linguistic context, and examples), lexicography and neologisms, and psycholinguistics and neologisms, examining how new words become entrenched and institutionalised. Within the lexicography segment, students explored the decision-making process used by Duden to include new words in the dictionary, including the role of corpus-based analysis. They were subsequently presented with a selection of neologisms and, working in groups, evaluated whether each term should be added to Duden, based on the editorial criteria. As frequency of occurrence is one of the key selection criteria, students then chose one neologism and conducted a preliminary frequency analysis using DeReKo.

Before engaging in independent corpus analysis, students were introduced to DeReKo and COSMAS II through a step-by-step registration process, guided by the lecturer. Since certain aspects of the platform were not immediately intuitive, this support ensured that all students could access the corpus without difficulty. Once registered, students were introduced to the different corpus archives and were informed that the W – Archiv für geschriebene Sprache would be used for research purposes. They then conducted a frequency analysis on a given neologism alongside the lecturer, following a modelling approach to develop familiarity with the tool. During this phase, they were also made aware of key considerations in corpus searches, including query formulation, data reliability, and data presentation.

Following this initial exercise, students transferred the corpus data into Excel and created visual representations of the results, again in parallel with the lecturer. They then compared their figures to those produced by the lecturer to refine their analytical skills. Once confident in their abilities, students conducted their first independent frequency analysis on a previously discussed neologism, with the lecturer providing guidance as needed.

The second phase of the course introduced the project, outlining its research questions, hypotheses (cf. Section 2.1), methodological steps, and group organisation. The project aimed to answer two research questions:

Two hypotheses were formulated:

Working in small groups (three to four students), participants selected ten COVID-related neologisms – five already listed in the online Duden and five not. They used the DWDS Themenglossar zur COVID-19 Pandemie5 and OWID Neuer Wortschatz rund um die Coronapandemie6 to identify relevant terms. Each group then conducted frequency analyses (2020-2023) in COSMAS II and created visualisations using Excel. They subsequently performed a co-occurrence analysis, interpreting and presenting their data autonomously but being offered support from the lecturer, when needed.

The first cohort (2022/23) delivered their findings through in-class presentations, while the second cohort (2023/24) created short video reports7. Presentations could be given in either German or Italian, depending on the students’ proficiency levels. The project concluded with a short critical written discussion of their own results for the 2022/23 cohort, and a written critical review of another video for the 2023/24 cohort.

At the conclusion of the course, students were invited to complete a feedback questionnaire to provide insights into their perceptions of the course. While the original questionnaire included questions on both the theoretical and practical components, the results presented here pertain exclusively to the practical aspect of the corpus project.

In 2022/23, 19 of the participating students completed the feedback form, while in 2023/24, 15 students did so. Initially, students were asked which part of the course they found most instructive – the theoretical or the practical component – and were requested to justify their choice. Of the 34 students who provided feedback across both years, the majority (20 students) considered the practical component to be more instructive, while seven students favoured the theoretical part, and seven found both components equally beneficial.

Students particularly appreciated that the practical component enabled them to apply the theoretical knowledge they had acquired, affording them a more hands-on experience that facilitated a deeper understanding of the subject matter (see student responses 1-3). Additionally, they valued the interactive and collaborative nature of the practical tasks (see student responses 4-5), as well as the opportunity to develop new practical skills (see student responses 6-7).

Students were subsequently asked to respond to a series of five-point Likert-scale questions regarding the practical component, with 5 indicating strong agreement and 1 indicating strong disagreement. These questions aimed to assess how demanding, motivating, engaging, and useful students found this aspect of the course. As illustrated in Figure 1, students generally perceived the practical component as demanding and challenging – 14 out of 34 students (strongly) agreed, 16 remained neutral, and 4 (strongly) disagreed. However, the majority found it motivating (24 out of 34 (strongly) agreed, with only 2 disagreeing), useful (23 out of 34 (strongly) agreed, with only 4 (strongly) disagreeing), and interesting (26 out of 34 (strongly) agreed, with only 1 disagreeing).

As the corpus project was designed not only to enhance students’ transferable skills and motivation but also to support their language development, students were asked whether working with the corpus had contributed to their metalinguistic knowledge and practical German skills. While the 2022/23 cohort was asked a single yes-or-no question on whether the practical component had improved their metalinguistic awareness and German proficiency, the questionnaire for the 2023/24 cohort was refined to include two separate five-point Likert-scale questions: one concerning metalinguistic knowledge and the other concerning language proficiency.

In the first cohort, 17 out of 19 students reported that working with the corpus had enhanced their theoretical and practical German skills. In the second cohort, 10 out of 15 students (strongly) agreed that the project contributed to their metalinguistic knowledge (with 5 remaining neutral), while 7 out of 15 (strongly) agreed that it had improved their German proficiency, 7 were neutral, and 1 disagreed.

Students were asked to elaborate on how the practical component had contributed to their linguistic development. Many reported improvements in their vocabulary (see student responses 8-10) and highlighted that working with the corpus allowed them to engage with authentic German usage (see student responses 10-11).

However, not all students perceived a tangible improvement in their practical German skills, particularly given that Italian was the dominant language in the classroom (see student response 12).

In line with this, one student suggested in the feedback section, “I would increase the use of German in the course”. Meanwhile, another student commented, “Simplify the materials provided, especially the videos”, – all of which were in German. This contrasted yet with another remark: “Reduce the number of times theoretical videos are watched in class”. These responses indicate differing preferences among students regarding the balance of German and Italian in instruction. While the majority of students found the corpus project beneficial in developing various skills, particularly metalinguistic knowledge (which is the primary objective of the course), only half of the 2023/24 cohort reported a perceived improvement in their practical German proficiency.