The importance of raising teachers’ and students’ awareness of pragmatics in German second language writing: a study of the effect of grammatical and lexical errors compared to pragma-linguistic infelicities

Abstract

One of the new scales in the CEFR Companion Volume is online interaction. The new descriptors cover goal-oriented online transactions, including written correspondence and the use of formulaic language. In Dutch business schools, students learn German for special purposes at B1/2 level: they are expected to master German in a professional context well enough to facilitate interactions and transactions. The question is how their not always flawless attempts at inviting a German colleague to a meeting are received by L1 language users. Is the level of B1/2 German sufficient to avoid bothersome situations? Is these learners’ mastery at B1/2 level sufficient to maintain a good relationship with a German partner? This study investigated whether communicative clashes might occur because of flawed German language use. Ninety-eight German business professionals, all L1 speakers of German, were asked to rate pragma-linguistic infelicities and syntactic, lexical and morphological errors in 16 business e-mails written by Dutch learners of German, using a 7-point Likert scale. The results showed that pragma-linguistic infelicities were considered more bothersome than syntactic, lexical or morphological errors, which suggests that L2 German writing courses should raise more explicit awareness of pragma-linguistic forces in intercultural communication. This idea is in line with recommendations contained in the new Companion Volume. Furthermore, it was studied whether knowledge of the writer’s foreign provenance and Dutch nationality affected the judgement of the German professionals. This was not the case.

Keywords:

• Second language acquisition
• error
• pragma-linguistic infelicity
• German
• written interaction

Introduction

Nowadays, it is often assumed that LX speakers (Dewaele, 2017) of English have to learn only one foreign language to be successful in intercultural business encounters, and that language is English. Studies of the language used in the intensive business interdependence between Germany and the Netherlands (trading volume €172 billion in 2020) reveal a different picture. Many German business people do not master English as well as their Dutch counterparts do (Busse, 2017; Eurobarometer, 2012). Therefore, German is used more often than English in Dutch–German business encounters (Vollstedt, 2005). German and Dutch belong to the same language family (Harding & Sokal, 1988) and share many features, and Dutch students in secondary schools undergo 5 years of compulsory German classes. Despite all this, communication in German still poses certain difficulties for Dutch business partners. For example, even advanced Dutch learners of German have difficulties using modal verbs correctly (Jentges, 2016), which might lead to misunderstandings. Earlier research has shown that errors in a first language negatively affect the reader’s attitude towards the writer (Beason, 2001; Jansen & Janssen, 2016; Martin‐Lacroux & Lacroux, 2017) and even the behaviour of the reader. Martin-Lacroux and Lacroux (2017) found, for example, that errors in application forms were a reason for recruiters not to select an applicant.

Planken et al. (2018) showed that errors in LX also had a negative impact on the L1 reader. They investigated the effect of LX writing (British English) errors on L1 and LX readers of British English. Their results show that identification of errors in an LX text not only influences readers’ evaluation of the text, but also negatively colours readers’ perceptions about an LX writer. This raises the question as to how errors made by Dutch learners of German are perceived by L1 German professionals.

In the Netherlands, classes in German writing usually focus on grammatical correctness. Less attention is paid to pragmatic inappropriateness, and to the pragmatic pitfalls for the language learner (Luijkx et al., 2020). The most recent edition of the CEFR Companion Volume (2018) explicitly calls for attention to pluricultural competences, and expounds further on the sociolinguistic and pragmatic competences already elaborated in the earlier edition, mostly with regard to spoken interaction. However, the curricula at the universities of applied sciences in the Netherlands have not yet been adapted to the revised framework. The focus on grammatical and orthographical accuracy in German language classes might also be inspired by a somewhat stereotypical image of German speakers being more concerned with linguistic flawlessness than Dutch language users. What is more, Dutch second-language teachers typically strive in their classes, to achieve an idealised ‘near-native’ level in their pupils, and flawless mastery of grammatical and pronunciation rules is at the heart of their teaching (see Piggott, 2019; Plante & Meester, 2019; Tajeddin et al., 2018).

Although several studies have investigated the effect of mistakes and errors of LX speakers on L1 speakers of American or UK English, it has never been tested whether the prototypical meticulousness ascribed to L1 speakers of German, and German professionals and business people in particular, corresponds to reality Just how disturbed are these individuals by mistakes, errors and infelicities made by Dutch learners of German in written communication?

Since the Netherlands and Germany share an intense relationship based on historical grounds and economic and political ties (Pekelder, 2014), it is possible that German readers might judge errors differently when they know the writer is Dutch. Van Oudenhoven et al. (2010) showed that German people hold a fairly positive attitude towards the Netherlands. This kind of goodwill factor may positively affect positively their judgement of errors Dutch people make in German. Rubin and Williams-James (1997) suggested that knowledge of the nationality of the writer may lead to judging errors in texts more leniently, although readers may differ as to the degree of awkwardness they perceive (Beason, 2001).

The results of our study may inspire authors of handbooks for Dutch–German interactions, especially in a business context, to implement the advice of the revised CEFR Companion Volume (2018) in their publications by paying attention to pluricultural and plurilingual techniques and featuring pragmatic differences more explicitly. This may help teachers and learners to become more aware of the errors and infelicities that should be avoided in Dutch–German interactions.

Errors, mistakes, infelicities and inappropriateness

The study of errors and inappropriateness is an important area of research in second-language acquisition. Studies in applied linguistics (for example: Brown, 2014; Ellis, 2015; Gass & Selinker, 2008; James, 2013) have shown the importance and value of analysing language learners’ errors. According to Ellis, an error is an unintended deviation from the rules of a target language or language variety made by a second-language learner. Such errors result from the learner’s ignorance of the correct rule (Ellis, 1994). According to Brown (2014), an error reflects the interlanguage competence of the learner.

Pragmatics addresses the appropriateness of utterances given specific situations, speakers and content. Pragmatic knowledge includes pragma-linguistic and socio-pragmatic knowledge. The distinction was famously introduced by Leech (1983) and Thomas (1983). Pragma-linguistics concerns the study of the resources that a language provides for meaning, whereas socio-pragmatics relates to situational context (Marmaridou, 2011; Sperlich et al., 2021). Pragmatic infelicities occur when LX language users unconsciously fail to correctly use the accepted way of using words and sentences in a target language, and these infelicities may lead to serious communication breakdowns (Cruz, 2013; House, 2012). As Gass and Selinker (2008) put it:

When breakdowns occur, they are frequently disruptive because L1 speakers attribute not linguistic causes [.], but personality (individual or cultural) causes. [.] Learners are generally unaware of this aspect of language and may be equally unaware of the negative perceptions that L1 speakers may have of them as a result of their pragmatic errors. (Gass & Selinker, 2008, p. 244)

Pragma-linguistic infelicities concern the appropriateness of utterances in specific situations by specific speakers and regarding certain content (Bardovi-Harlig et al., 1998). A pragma-linguistic infelicity is not wrong in grammatical terms, but fails to achieve the speaker’s goal (Thomas, 1983). Socio-pragmatic inappropriateness reflects the learners’ misjudgement of the context and situation in which linguistic resources are implemented (Taguchi, 2015).

On several occasions, Bardovi-Harlig and co-authors have highlighted that pragmatic awareness is not always sufficient to achieve the appropriate degree of pragmatic adequacy. Barcovi-Harlig has shown that limited linguistic proficiency may hamper the expression of acceptable pragmatic intentions. Pragmatic subtleties often can only be expressed when the linguistic repertoire is sufficient (Bardovi-Harlig, 2017).

In the revised edition of the CEFR Companion Volume (2018), the scale of ‘online interaction’ was included for the first time. For the intermediate level (B1-B2), the ‘can-do’ statements that describe the proficiency level state the following: ‘Can use formality and conventions appropriate to the context when writing personal and professional letters and e-mails. Can write formal e-mails/letters of invitation, thanks or apology with appropriate register and conventions. Can write non-routine professional letters, using appropriate structure and conventions, provided these are restricted to matters of fact. Can obtain, by letter or e-mail, information required for a particular purpose, collate it and forward it by mail to other people’ (CEFR Companion Volume, 2018, p. 94).

It can be deduced that the morphological errors at B1/B2 level concern the declension or conjugation of nouns, verbs, and adjectives. Syntactic errors concern structures that are larger than the word, namely phrases, clauses, and sentences. Lexical errors concern the incorrect usage of a word or fixed expression.

Pragma-linguistic infelicities that Dutch L2 learners at the B1/B2 level of German are prone to make concern the use of address terms and modal verbs and choices about the degree of informality (for a discussion of possible language features that signal pragma-linguistic inappropriateness, see Bardovi-Harlig, 1999).

Effect on the L1 recipient

Research into the effect of linguistic errors or pragma-linguistic infelicities on the L1 language user is rare, especially for German language users. Most of the current literature on LX error perception is about the effect of grammatical errors on the reader (see Rifkin & Roberts, 1995, for an overview), and only a few studies address the effect of pragma-linguistic infelicities (Aimoldina et al., 2016; Hendriks, 2010; Zhu, 2012).

The majority of studies explore the effect of spoken or written errors in L2 English. Schauer (2006), for example, studied awareness of grammatical errors versus pragma-linguistic infelicities in L2 spoken English. L2 English (ESL) learners in Germany, ESL learners in the UK, and English L1 speakers in the UK were asked to rate spoken fragments containing grammatical errors and pragma-linguistic infelicities. The ESL learners in the UK and the L1 speakers were more aware of the pragma-linguistic infelicities than the ESL learners of English in Germany. This confirmed the earlier conclusions of Bardovi-Harlig et al. (1998). Subsequently, Schauer interviewed all three groups to explore which errors and infelicities they found most bothersome. Sixteen of the twenty L1 speakers of English (80%) stated that pragma-linguistic infelicities were more bothersome than grammatical errors.

Wolfe et al. (2016) explored the impact of grammatical errors and pragma-linguistic infelicities in written American English on L1 speakers of American English in a business setting. They selected errors and infelicities that, although noticeable, did not substantially interfere with comprehension. Participants were asked to comment on three versions of an e-mail. The first of these contained grammatical errors by an L1 writer, the second contained grammatical errors by an LX writer and a third e-mail contained pragma-linguistic infelicities. For this third e-mail only, the participants did not know whether the writer was an L1 speaker of American English. The study’s results showed that pragma-linguistic infelicities of politeness and tone were perceived as more bothersome than grammatical errors. Furthermore, they found that participants made allowances for grammatical errors made by LXs. Because no distinction was made between an L1 and an LX writer in the e-mail with pragma-linguistic infelicities, we do not know whether pragma-linguistic infelicities committed by L1s are judged differently than pragma-linguistic infelicities made by LXs. These results of Wolfe et al. (2016) are in line with those of Bardovi-Harlig et al. (1998) and Schauer (2006) on the effect of L2 English on L1 speakers of English. All three studies found that L1 speakers of English consider pragma-linguistic infelicities to be more bothersome than grammatical errors.

Hasbún (2001) explored how apologies and requests written in English by L1 speakers of Spanish were perceived by L1 and LX teachers of English. L1 speakers and LX speakers had to rate the pragma-linguistic appropriateness and severity of grammatical errors in these texts. The results indicated that L1 speaking teachers ranked pragma-linguistic errors as more serious than grammatical errors. LX teachers of English, however, showed the opposite pattern: they ranked grammatical errors as more serious than pragma-linguistic infelicities. These findings are in line with those of Bardovi-Harlig et al. (1998) and Schauer (2006).

Delisle (1982) evaluated the effect of written errors made by learners of German on speakers of German. Her study measured the effect of different errors in morphology (gender, verbal morphology and case endings), syntax (word order), vocabulary and spelling. In her study, 193 German pupils aged 10–17 years were asked to choose which of two errors was worse. The pupils ranked gender errors as the worst kind of error, and spelling errors as the least serious. Unfortunately, Delisle did not include pragma-linguistic infelicities in her study. It is unclear whether the results of the evaluation also hold for adult L1 speakers of German.

A number of other studies highlight the type of grammatical errors and pragma-linguistic infelicities that occur in L2 German. Bohnacker and Rosén (2007), for example, showed that Swedish learners of German used Swedish word order in German verb-second (V2) declaratives. Jentges (2016) found that very advanced Dutch learners of German struggled constantly with the use of modal verbs. They tended to use the indicative form of sollen instead of the subjunctive; e.g. Sie soll das tun (She is supposed to do this) instead of Sie sollte das tun (She should do this). This might impact negatively on the sender because, to the German-speaking receiver, soll comes across as more direct than sollte does.

Multiple studies have investigated the acquisition of specific difficult pragma-linguistic aspects in German, such as address terms (Kinginger & Belz, 2005), modal particles such as eben or doch (Belz & Vyatkina, 2005; Möllering, 2001), and modal verbs (Cunningham & Vyatkina, 2012). All these studies focus on learning pragma-linguistic aspects of spoken German. A good example is the study by Cunningham and Vyatkina (2012), where American students learned specific formal pragma-linguistic aspects by practicing modal verbs in online conversations with German professionals or with peers in Germany. To our knowledge, our study is the first to contrast socio-linguistic infelicities with syntactic, lexical or morphological errors in German writing.

Knowing that the sender is foreign or knowing that the sender is Dutch

Kecskes (2015) shows that the discursive turn in intercultural impoliteness implies that awareness or knowledge of the pragmatic situation, such as knowing that a speaker is LX or has a certain cultural background different from that of the L1 receiver, impacts on the effect of the possible error or infelicity (Kecskes, 2015). When language users mutually understand that they do not share a mother tongue, they may infer that the literal meaning is not intended or those messages are intended otherwise than how they were initially perceived. Apparently, language users have different norms for LX speakers, and they may show some leniency with regard to linguistic errors, but also with regard to pragmatic inappropriateness. Therefore, it may be the case that if L1 language users know that the writer of a message containing inaccurate or inappropriate language use is not a L1 language user, their reaction or attitude may be different.

Apart from the fact that one can be lenient towards errors and infelicities made by LXs in general, it is also possible that one is especially lenient towards LX speakers from a certain country as a result of historical or economic ties or the geographical situation. Only a few studies have investigated whether knowledge of a writer’s nationality influences the judgement of the L1 speaker. Rubin and Williams-James (1997) found that American instructors who knew the nationality of their students were more lenient with errors and infelicities in essays written by Danish and Thai students than with similar errors and infelicities in essays written by American students. This result is in line with the study by Wolfe et al. (2016) discussed above.

Van Oudenhoven et al. (2010) investigated the degree of affinity between Dutch and German people on the basis of social identity theory (Tajfel, 1978; Tajfel & Turner, 1979). Their theory assumes that people strive to maintain or achieve a positive social identity. One’s own group is compared with other groups, and based on positive outcome of those comparisons, a positive social identity is formed. Van Oudenhoven et al. (2010) applied this theory to the attitudes of small and large nations towards each other: Belgium, the Netherlands and Denmark (the small nations) versus Germany, the UK and France (the large nations). Their results show that German participants rated the small nation of the Netherlands as fairly likable. As social identity theory predicts, they did not feel threatened by the small country. This positive attitude could influence the outcome of our study in the sense that L1 Germans might perceive pragma-linguistic infelicities and formal errors to be less bothersome when they know the writer is Dutch.

The results of our study may inspire authors of handbooks for Dutch–German interactions, especially in a business context, to implement the advice of the revised CEFR Companion Volume (2018) in their publications by paying attention to pluricultural and plurilingual techniques and featuring pragmatic differences more explicitly. This may help teachers and learners to become more aware of the errors and infelicities that should be avoided in Dutch–German interactions.

Study purpose and hypotheses

In this article, we focus on linguistic errors and pragma-linguistic infelicities of Dutch learners on the receiving end of the communication: L1 German professionals and business people. More specifically, we asked our respondents how ‘bothered’ they were by a certain error or infelicity, that is, the extent to which being confronted with a specific error or infelicity caused them to experience feelings of awkwardness, annoyance or irritation. We purposely use the term ‘bothered’, because it is the translation of the German ‘störend’ that we used in the questionnaire. Störend is commonly used in German to express feelings of unease, discomfort or slight irritation that may occur when reading something that is not completely accurate or desirable. The errors and infelicities whose störend (bothersome) qualities the respondents had to rate were printed in boldface so that we could be quite sure only these errors and infelicities were judged, and no other aspects of the text. Since Bardovi-Harlig et al. (1998), Hasbún (2001), Schauer (2006) and Wolfe et al. (2016) found that L1 speakers of English judged pragma-linguistic infelicities more negatively than grammatical errors, our first hypothesis was:

H1: German readers’ judgements will be affected more negatively by pragma-linguistic infelicities than by syntactic, lexical or morphological errors.

It appears from the literature that issues other than error or infelicity might affect a reader’s evaluation. This includes degree of affinity with the writer’s nationality. Van Oudenhoven et al. (2010) showed that German participants perceived the Netherlands to be a likable nation. On the basis of this outcome, our second hypothesis was:

H2: German readers’ judgements of pragma-linguistic infelicities and morphological, syntactic, or lexical errors are less negative when they know the writer is Dutch than when they do not know the writer’s nationality.

Method

We developed a quasi-experiment with a within-subject design that asked L1 German speakers to evaluate four versions of a short e-mail.

Stimuli

The stimuli in this study consisted of e-mails with pragma-linguistic infelicities and morphological, syntactic and lexical errors. In order to ensure the best possible ecological validity of our data, the morphological, lexical and syntactic errors were gathered from a B2 student error database. This was a sample of e-mails written by 25 business communication students at a university of applied sciences in the Netherlands. In total there were 16 short e-mails: four e-mails containing four different pragma-linguistic infelicities, four mails containing four different morphological errors, four mails containing four different lexical errors and four containing four different syntactic errors. Table 1 presents examples of the errors and infelicities; See Figure 1 in the results section for an example; Appendix A gives a complete list of all errors and infelicities in the stimuli.

Figure 1. Example of an E-mail with errors and infelicities in bold.

Table 1. Examples of errors and infelicities in the e-mails.

Category	Incorrect form (in bold)	Correct form
Morphological error	mit die Hochschule	mit der Hochschule
Syntactic error	wenn Sie neugierig sind geworden	wenn Sie neugierig geworden sind
Lexical error	auf dem 11. Juni	am 11. Juni
Pragma-linguistic infelicity	Hey Frau Müller	Sehr geehrte Frau Müller

In the four e-mails with pragma-linguistic infelicities, each e-mail contained one infelicity in the salutation, one infelicity in modal verb usage, one infelicity in address terms (e.g. du instead of Sie), and one infelicity in the letter closing. See Example 1 for an e-mail with pragma-linguistic infelicities.

The errors and infelicities were placed in a context that was kept similar throughout the survey: a request in the form of a business invitation. This context was chosen in order to create the best possible ecologically valid stimulus ( based on Beason, 2001; Planken et al., 2018; Roberts & Cimasko, 2008). We used the format of a business invitation to take part in a brainstorming session or network, and invited the reader to attend an evening information session.

In other research, the errors and infelicities were left implicit, and they often went unnoticed by the reader (Planken et al., 2018). In order to prevent this from happening and to be relatively sure that a judgement could be related to an error or infelicity, we explicitly highlighted the errors and infelicities, following Beason (2001). All errors and infelicities were printed in bold.

We used a Latin square design to make sure all participants randomly judged all four categories of errors/infelicities (morphological, syntactic, lexical and pragma-linguistic).

Participants

Participants were recruited via German university networks, networks on Xing (the German version of LinkedIn) and German Rotary Club networks. To avoid bias based on the country of origin, we created a German e-mail address. We sent e-mails to German contact persons working in a business environment, and asked them to distribute the link with the online questionnaire to other professionals. All the participants were informed beforehand that the questionnaire was about evaluating e-mails written by business people from another country, but they did not know that the purpose of the research was to find out how bothered they were by infelicities and errors in German. Participants were assured that anonymity was guaranteed.

The 98 participants were speakers of German who worked in a business environment for either the government or a non-governmental organisation. All stated that their mother tongue was German, and they could accordingly be considered L1 speakers of German. We chose these participants because we believed they would be the first to receive an invitation e-mail of the kind we composed in real life. They would also be in a position to make decisions about further collaboration with a Dutch company.

Of the participants, 48% were male and 52% female. 70% had a university degree (Bachelor’s, Master’s, Doctorate) and the other 30% had less education. The average age of the participants was 49 years (range: 20–69 years). To check whether older people judged errors differently from younger people (Vann et al., 1984), we formed two age groups: 20–45 and 46–69. Independent t- tests found no significant difference in the judgements on the morphological, syntactic, and lexical errors or the pragma-linguistic infelicities between the female and male participants, the two levels of education or the two age groups. Therefore, we did not introduce the independent variables gender, education and age in our analyses.

Procedure

We did not reveal the writer’s nationality until the respondent started the questionnaire. Half of the participants were randomly shown introduction A, which explained that the e-mails were composed by business people from another country. The other half read introduction B, which explained that the writer was Dutch (see Appendix B for the introductions). This enabled us to compare the judgements of a group whose expectations about the Dutch were activated before they read the e-mails to a second group who judged the e-mails without a Dutch writer in mind (Hypothesis 2). It took the participants 12 min on average to fill out the questionnaire. Once the survey was completed, participants were informed about the purpose of the experiment, about the Dutch nationality of the research group, and about the fact that the survey was hosted by Radboud University.

Measuring instruments

Participants were asked to rate the extent to which ‘This error bothers me’ (based on Beason, 2001) on a 7-point Likert scale (1 = trifft vol und ganz zu (strongly agree), 2 = trifft zu (agree), 3 = rifft eher zu (somewhat agree), 4 = neutral, 5 = trifft eher nicht zu (somewhat disagree), 6 = trifft nicht zu (disagree), 7 = trifft gar nicht zu (strongly disagree). See Study Purpose and Hypotheses for an explanation of the term ‘bothersome’.

Some German readers might recognise typically Dutch errors or infelicities, and would therefore judge the e-mails with a Dutch writer in mind. To check whether the readers of introduction A, which did not inform participants that the writer was Dutch, had a certain nationality in mind whilst judging the language errors, they were asked: ‘What country do you think the writer comes from’?

Multiple-choice and open-ended questions were used to collect background data on age, gender and highest educational level.

Statistical methods

In order to check whether it would be possible to summarise the data for each of the four errors/infelicities, we first performed reliability analyses by calculating Cronbach’s alpha. All categories were α > 0.70: lexical (α = 0.80), syntactic (α = 0.81), morphological (α = 0.90) and α = 0.74 for the pragma-linguistic category in the four e-mails. The high alphas showed that it was possible to summarise the data for each of the four categories. In order to check whether the four e-mails with the same category of errors/infelicities were judged equally bothersome, we subsequently performed four one-way ANOVAs to test whether there was a significant difference in rating between the four e-mails with the same error/infelicity category. No significant differences in ratings were found among the four e-mails with morphological errors (F(3.94) = 0.27, p = .847, η² = 0.009), the four e-mails with syntactic errors (F(3.94) = 0.19, p = .903, η² = 0.006) and the four e-mails with lexical errors (F(3.94) = 0.69, p = .561, η² = 0.022). However, a significant difference was found among the four e-mails with pragma-linguistic infelicities (F(3.94) = 26.52, p<.001, η² = 0.458). E-mail 15 was judged to be less bothersome than the other three e-mails with pragma-linguistic infelicities. A comparison of the judgement of the infelicities in e-mail 15 (M = 5.03, SD=1.54) with the other three e-mails with pragma-linguistic infelicities suggests that e-mail 15 deviated from the other e-mails in this category in the sense that it did not show a register clash whereas the other three did. We, therefore, decided to exclude e-mail 15 from the analysis. We will discuss this e-mail in the conclusion and discussion section and in the limitations and further research section. The deletion of the results for e-mail 15 resulted in a reduction of the participants to 69, since within-subject analyses were the most suitable analyses to test our hypotheses.

To test the first hypothesis, we performed a repeated measure ANOVA to examine whether there were significant differences in ratings across the four error categories of pragma-linguistic, lexical, morphological, and syntactic. For this analysis we used the summarised data per category: the 16 morphological errors in the four e-mails with morphological errors, and so on. For the pragma-linguistic infelicities, we left the four infelicities in e-mail 15 out of consideration. To test the second hypothesis, we performed a 2 by 4 mixed ANOVA on ratings with error category (pragma-linguistic, morphological, syntactic, and lexical) as a within-subject factor and knowledge of the writer’s nationality (no knowledge of nationality of L2 writer, knowledge of Dutch nationality) as a between-subject factor.

Results

Table 2 shows the mean ratings per category (the boldfaced results) after we removed e-mail 15. A repeated measures ANOVA showed that there was a significant difference within the means of the four categories (F (2.71, 184.14) = 11.89, p < .001, partial η² = 0.149). Mauchly’s test indicated that the assumption of sphericity had been violated, χ²(5) = 11.16, p = .048; therefore, Greenhouse Geisser (ε = 0.90) corrected tests were reported. Post-hoc comparisons demonstrated that the pragma-linguistic infelicities were significantly more bothersome than the lexical (t(68) = 5.418, p < .001, d= 1.59), the morphological (t(68) = 4.162, p = .001, d = 1.89) and the syntactic errors (t(68) = 5.216, p<.001, d = 1.40).

Table 2. Mean rating of morphological, syntactic and lexical errors and pragma-linguistic infelicities with knowledge of Dutch nationality and without knowledge of nationality.

Category of error/infelicity		M	SD	n
Pragma-linguistic infelicity	Knowledge of nationality	2.61	1.02	29
	No knowledge of nationality	2.44	1.43	40
	Total pragma-linguistic infelicities	2.52	1.23	69
Morphological errors	Knowledge of nationality	3.66	1.48	29
	No knowledge of nationality	3.32	1.88	40
	Total morphological errors	3.46	1.72	69
Syntactic errors	Knowledge of nationality	3.16	1.11	29
	No knowledge of nationality	3.56	1.47	40
	Total syntactic errors	3.39	1.34	69
Lexical errors	Knowledge of nationality	3.70	1.42	29
	No knowledge of nationality	3.45	1.63	40
	Total lexical errors	3.55	1.54	69
	Total knowledge of nationality	3.28	.88	29
	Total no knowledge of nationality	3.19	1.22	40

Note: Likert scale values: extremely bothersome (1), not bothersome at all (7). Total numbers are in bold.

Table 2 also presents the results of the second hypothesis: German readers’ judgements of pragma-linguistic infelicities and morphological, syntactic and lexical errors are less negative when they know the writer is Dutch than when they do not know the writer’s nationality. The results showed a significant main effect of Category (F(2.66, 178.37) = 11.62, p<.001, partial η² = 0.153) which was similar to the effect reported above, and a non-significant main effect of Knowledge of nationality (F(1) = 0.11, p = .747, partial η² = 0.002). No significant difference in judgement of pragma-linguistic infelicities and morphological, syntactic, and lexical errors (F(1,67) = 1.58, p = .214, partial η² = 0.023) was found between participants who were informed the writer was Dutch and participants who were not.

The participants who read introduction A (stating that the letters were written by foreign writers) were asked: ‘Which country do you think the writer comes from’?. Answers varied greatly: 35% said the Netherlands, 13% said Eastern Europe, 11% the UK, 8% France, 6% Sweden, 13% India, Turkey, Spain, Belgium, Luxemburg or Germany and 14% did not know. It is notable that 35% of the participants who were not informed about the writer’s nationality correctly surmised the writer was Dutch. This could mean they recognised typical Dutch errors or failures, and therefore, judged them with a Dutch writer in mind. Because this could have influenced our findings, we performed a 2 by 4 mixed-design ANOVA on judgement with Category (morphological, syntactic, lexical error or pragma-linguistic infelicity) as a within-subject factor and thinking the writer was Dutch (n = 14) as opposed to thinking the writer had another nationality (n = 26) as a between-subject factor. Mauchly’s test indicated that the assumption of sphericity had been violated (χ²(5) = 16.85, p = .005); therefore, Greenhouse Geisser (ε = 0.78) corrected tests were reported. The results showed a significant main effect of Category (F(2.35, 89.10) = 5.83, p = .003, partial η² = 0.133) similar to the effect reported above, and a non-significant main effect of the Dutch versus other nationality groups (F(1, 38) = 0.07, p = .800, partial η² = 0.002). No significant differences between participants who thought the writer was Dutch and respondents who thought the writer had another nationality were found in judgement of pragma-linguistic infelicities and morphological, syntactic and lexical errors (F(2.35, 89,10)=0.40, p = .709, partial η² = 0.010).

We then performed a 2 by 4 mixed ANOVA on judgement with pragma-linguistic infelicities and morphological, syntactic and lexical errors as a within-subject factor and two new groups. The first group comprised the 14 participants that surmised the writer was Dutch plus the 29 participants who were informed in the introduction that the writer was Dutch (n = 43), while the second group comprised the participants who were not informed in the introduction that the writer was Dutch minus the 14 that surmised the writer was Dutch (n = 26) as a between-subject factor. The results showed a significant main effect of Category (F(3,201) = 11.77, p < .001, partial η² = 0.149) similar to the effect reported above, and a non-significant main effect of the two new groups (F(1,67) = 0.00, p = .998, partial η² = 0.000). No significant differences in judgement of pragma-linguistic infelicities and morphological, syntactic, and lexical errors were found between the participants who knew or guessed the writer was Dutch and those who did not.

Discussion and conclusion

In order to find out what kinds of error types or infelicities may cause bothersome situations in German- Dutch business encounters, we investigated which category bothered L1-German business professionals in business e-mails most: pragma-linguistic infelicities or morphological, syntactic, and lexical errors. Our findings confirm our first hypothesis: German readers’ judgements are affected more negatively by pragma-linguistic infelicities than by morphological, syntactic, and lexical errors. These outcomes corroborate the studies of Hasbún (2001) and Wolfe et al. (2016) on written L2 English, and those of Bardovi-Harlig et al. (1998) and Schauer (2006) on spoken L2 English. We thus see a pattern for the acceptability of LX errors and infelicities in written German similar to that which was established in written and spoken English. Our results are in line with the updated edition of the CEFR Companion Volume (2018), which addresses formal written correspondence for the first time, stressing pragma-linguistic considerations such as the importance of adequate professional correspondence and the appropriate use of tone, style and formulaic expressions (p. 94). Our study suggests that Dutch teachers of German and textbooks for teaching German should pay more attention to the pragma-linguistic aspects we investigated, because their impact exceeds that of the morphological, syntactic, and lexical errors we tested.

As discussed above, e-mail 15 was left out of consideration because it was judged significantly less bothersome than the other e-mails with pragma-linguistic infelicities. A closer look at e-mail 15 (see Example 2, right column) reveals that it was the only mail that did not exhibit pragma-linguistic inconsistency in terms of formality. E-mail 15 was a fairly consistent informal e-mail, whereas e-mail 13, for example, used the salutation Hey Frau Müller, which is a combination of the impolite and informal (Hey) and formal (Frau Müller). Could this mean that not only the terms of address play a part in the perception of a pragma-linguistic infelicity, but also the stylistic inconsistency of a message? This calls for further research. We will discuss this in more detail in the Limitations and Further Research section.

Our second hypothesis expected that knowledge of the writer’s nationality would influence the readers’ judgement of the errors and infelicities, but this was not the case. Despite the German people’s high degree of affinity with the Dutch people as reported by Van Oudenhoven et al. (2010), our data did not show that goodwill factors influenced participants’ assessments of their neighbors’ language use. A comment from one of the respondents illustrates the tendency to be lenient towards all LX writers. She said: ‘I believe one should be considerate and understanding if someone from another country is trying to speak your language’.

Limitations and further research

A limitation of our research is that we restricted ourselves to one speech act: we only worked with e-mail requests in the form of a business invitation. We did this in order to be able to ascribe our findings to the type of error or infelicity and to exclude possible other factors. Further research is necessary to reveal if our findings hold true for other speech acts as well. A second limitation is that the errors and infelicities were boldfaced in the e-mail so that the readers would not overlook them. This allowed us to measure exactly which error or infelicity bothered the reader most. However, in real life, errors and infelicities are not boldfaced, which makes it easier to overlook them. Future research on error and infelicity evaluations should compare highlighted errors with non-highlighted ones.

Sending a survey by e-mail to the participants allowed us to reach business people that generally have limited time to fill out surveys. However, a disadvantage of this method is that the researcher loses control of many aspects related to filling in the questionnaire (e.g. situation, motivation and concentration). In future research, one might consider asking a subset of the participants for a retrospective interview.

Another limitation is that we had to leave out e-mail 15. Because we left out e-mail 15, our design may seem rather unbalanced – 3 e-mails with pragma-linguistic infelicities and 4 e-mails for each morphological, syntactic and lexical category – but, as no significant difference was found between the e-mails with the morphological, syntactic and lexical errors, the interpretation of the outcome remains the same. E-mail 15 appeared to be an outlier in that it tested the reaction to a consistently informal e-mail; it contained no register clash. It reveals that in German business encounters, an informal approach for a foreign business partner might not necessarily be a deal-breaker for L1 speakers of German as assumed by Cunningham and Vyatkina (2012). They discuss the necessity of an appropriate degree of formality. We, on the other hand, point out that register clash or register inconsistency has, to our knowledge, never been qualified as a typical pragma-linguistic infelicity. Our research has shown that, in the case of LX learners, inconsistency in register seems to be more bothersome than an inappropriate degree of formality. The ratings thus showed that our German participants did not find such a consistently informal e-mail very bothersome. Further research should shed more light on this topic.

Our study shows that breaking pragma-linguistic conventions within Dutch–German business encounters might be a cause for friction, and that their impact exceeds that of the morphological, syntactic and lexical errors we tested. We therefore recommend that Dutch teachers of German should pay more attention to pragma-linguistic infelicities and teach their students how to avoid and amend frictions caused by such infelicities. Many studies like Bardovi-Harlig (2017), Cheng (2016), French and Beaulieu (2016), Huth (2010), Li and Gao (2017), Liang (2021) and Taguchi (2015), offer suggestions on how to teach pragma-linguistics, for instance via metapragmatic discussions. Metapragmatic discussion occurs when students and teacher engage in a dialogue about what they said and why they said it. Still, there is little information on the extent to which these tools are incorporated in current teaching practices. Our future research will therefore analyse the current teaching materials for Dutch students learning German, assess the degree of attention paid to pragma-linguistics in these materials, and indicate whether fine-tuning is called for.

Open Scholarship

This article has earned the Center for Open Science badges for Open Data and Open Materials through Open Practices Disclosure. The data and materials are openly accessible at https://doi.org/10.34973/28t0-km68 and https://doi.org/10.34973/28t0-km68. To obtain the author’s disclosure form, please contact the Editor.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Antoinette Luijkx

Antoinette Luijkx studied German language and culture at Radboud University Nijmegen. As a lecturer, she teaches German to Dutch students at the University of Applied Sciences in ‘s-Hertogenbosch.

Marinel Gerritsen

Marinel Gerritsen (PhD: University of Leiden) is emeritus professor of Intercultural Business Communication Studies at Radboud University Nijmegen.

Margot van Mulken

Margot van Mulken (PhD: VU University Amsterdam) is a full professor in International Business Communication Studies at Radboud University Nijmegen.

Appendix A

Table

Category of error/infelicity	Example
Pragma-linguistic infelicities	Hey Frau Müller, Lieber Peter, Hallo Anja, Hey Herr Müller Sie und Ihre Kollegen sollen teilnehmen denn wir sollen eine Subvention mit Ihrer Hilfe beantragen. Anmelden muss hier! denn du wirst auch Vorteile Würdest du den Unternehmensplan Bist du daran interessiert Möchtest du teilnehmen Viele Grüße , Tschüß, Viele Grüße Gruß
Lexical errors	Workshops vor dem Erfahrungsaustausch eingeplant auf dem 11. und dem 12. Juni auf unsere Unternehmung Das meint, dass Sie erfahren werden Bei Interesse versuchen wir Sie das nötige Wissen verkriegen Das Projekt befördert die Kooperation die künftige Zusammenwirkung möglichst reibungslos verläuft ein geschicktes Datum die Anleitung unseres Schreibens ist
Syntactic errors	wenn Sie neugierig sind geworden der uns Ihr Unternehmen hat empfohlen der als Erster am Nachhaltigkeitsprojekt hat teilgenommen die Sie sich haben gewünscht Zusammen mit… Landschaftsleitung immer mehr …… Wir würden uns freuen Sie am 14. Juni zu dürfen begrüßen Wäre es nicht schön um alle zu können beitragen? die Kooperationen mit anderen Unternehmen können herstellen mit dem wir die Erstellung von einem Projekt können durchführen Für Rückfragen können Sie mich gerne erreichen unter die Nummer Möchten Sie mehr erfahren über unser Nachhaltigkeitsprojekt? Sie können mit den richtigen Partnern beschreiten neue Wege können wir antworten auf die ganz großen Herausforderungen
Morphological errors	mit die Hochschule zu ein Maximum mit ein Projekt u der Frühstück mit Sie In der Anlage finden Ihnen für Ihnen einen erste Informationsabend mit den andere Betrieben den wirtschaftliche Nutzen die regionale Unternehmen Sie sein Ich möcht wirden wir

Appendix B

Table

Instructions – German	English
Die Kommunikation mit den Unternehmen im Ausland erfolgt oft auf Deutsch. Man bemerkt es natürlich, wenn die deutsche Sprache des internationalen Handelspartners nicht perfekt ist. Uns interessiert Ihre Meinung zu bestimmten Unvollkommenheiten in der schriftlichen Kommunikation. Welche Unvollkommenheiten sollten besser nicht gemacht werden? Die Ergebnisse werden genutzt, um den Sprachunterricht der Fachhochschulen und Hochschulen in Europa zu verbessern.	Communication with foreign organisations often occurs in German. Naturally, it does not go unnoticed when the language of the international business partner is not perfect. We are interested in your opinion on certain imperfections in written communication. Which errors should not be made? The results will be used to improve language classes at European universities.
Die Handelsbeziehungen zwischen Deutschland und den Niederlanden. Deutschland ist schon seit Jahren mit Abstand der wichtigste Handelspartner der Niederlande, sowohl was Import als Export betrifft. Wenn Sie mit Niederländern kommunizieren, bemerken Sie natürlich, dass die deutsche Sprache der niederländischen Handelspartner nicht perfekt ist. Uns interessiert Ihre Meinung zu bestimmten Unvollkommenheiten in der schriftlichen Kommunikation. Welche Unvollkommenheiten sollten besser nicht gemacht werden?	The business relationship between Germany and the Netherlands For years now, Germany has been the Netherlands’ most important business partner for imports and exports. When communicating with the Dutch, you will of course that their German language use is not perfect. We are interested in your opinion on certain imperfections in written communication. Which errors should not be made?