From word to workplace. An exploratory research on language style and employee engagement.

(1)

From word to workplace

An exploratory research on language style and employee engagement

Martijn Rietveld

Master thesis S1285343

Supervisor: Dr. F. Jungslager

Second reader: Dr. M. van Leeuwen

MA Linguistics: Language and Communication Leiden University

10-5-2019

Abstract

This thesis explores the connection between language style and employee engagement. Employee engagement is the main objective of internal communication, a field which has remained largely unexplored from a stylistic point of view. In analyzing this connection between style and employee engagement, this exploratory study employs a combination of methods. A questionnaire analysis is conducted in order to achieve a reader’s impression on the employee engagement of five corpus texts. Then, a linguistically founded style analysis explores which stylistic features are likely to contribute to this reader’s impression. These analyses suggest patterns between four stylistic features and their contribution to employee engagement: use of concrete nouns, transitivity, tropes, and avoiding repetitive word usage.

(2)

Sometimes one word can make all the difference – a saying that John Lewis, a chain of department stores in the UK, has taken to heart. In an effort to keep workers for the company engaged, John Lewis refers to its workers not as ‘employees’, but rather as ‘partners’. By naming employees as partners, Oplift states that employees ‘are made to feel like their opinion matters right from the start. This empowers them to feel confident that they are valued by senior management and that their voice is heard’ (Lloyd, 2018).

Furthermore, The Guardian refers to the company as having an employee model that others could learn from (Wood, 12) and it is said that addressing employees as ‘partners’ has an engaging effect: the Chartered Management Institute (CMI) mentions how this invites employees to feel involved and to take personal responsibility on behalf of the company (Flint & Vinberg, 2015). Among other measures, the change of a single word lead to the company being lauded for its effort put into employee engagement.

Employee engagement is the main objective of internal communication, which is responsible for communication within an organization. By creating employee engagement, internal communication has a positive impact on aspects such as an employee’s

commitment or his ‘understanding of the organization’s evolving aims’ (Welch 2008: 495). Internal communication is a growing specialization (Vercic et al. 2012: 223), but

scholarship has not been able to keep pace with the rapid development of the initiatives from practitioners. As such, the definition of internal communication is varied and the field is not yet recognized as an independent domain (Vercic et al. 2012: 226). In order to create employee engagement through internal communication, Welch (2012: 253) stresses that applying persuasive writing skills is crucial, highlighting the importance of language and style within this field.

Language style can therefore play a significant role in employee engagement. We see this in the example of John Lewis: naming employees as ‘partners’ causes employees to feel more involved. These two terms are stylistic variants of the same definition (‘people working for John Lewis’) and they display a variation in semantics. This semantic variation

(6)

2 shows that stylistic variants can cause different effects (Van Leeuwen 2015: 14-15). In the case of John Lewis, addressing workers as ‘partners’ rather than employees has an engaging effect: it empowers employees, and engages them to feel personal responsibility for the company.

The example brought by John Lewis shows the effect of the stylistic variation of a single word. Renkema signifies that differences in morphology, syntax and composition all can bring various communicative effects (1987: 163). One way to uncover the potential effects of different stylistic phrasings is by conducting a stylistic analysis. Renkema states that when conducting a stylistic analysis, it is not merely about describing the differences in phrasing. Stylistic analysis instead focuses on elaborating these differences, analyzing the potential communicative effects these stylistic choices have (Renkema 1987: 163). The field of stylistics can help in clarifying why certain stylistic features are used in particular social contexts – and what communicative effects they bring within those contexts

(Renkema: 163).

In his dissertation Stijl en Politiek, Van Leeuwen (2015: 17) employs a linguistically founded stylistic research. This study analyzes how linguistic choices made in phrasing can contribute to the effect a text creates as a whole on the reader or listener (Van Leeuwen: 17). The linguistic style analysis indicates or demonstrates how certain stylistic choices are likely to cause certain communicative effects. For example, Van Leeuwen analyzes the speeches of Geert Wilders and Ella Vogelaar, known to have a ‘clear’ and ‘woolly’ manner of speaking respectively (2015: 40), and then studies how certain stylistic choices are likely to cause readers and listeners to have these impressions.

Van Leeuwen therefore focuses in his dissertation on the domain of political speeches, and a similar stylistic text analysis has predominantly been applied to literary contexts and some corporate contexts (Van Leeuwen 2015: 17). In this thesis, however, the insights a stylistic analysis can provide will be applied to the largely unexplored field of internal communication. As a social context that attempts to create employee engagement through informing, connecting and motivating employees by using persuasive and effective writing, internal communication is a field that can benefit from the insights provided by a

(7)

3 stylistic analysis. Linguistic research on the potential communicative effects of linguistic features is various and well established, but these communicative effects have rarely if ever been connected to the role of creating employee engagement. A style analysis can help in discovering how stylistic features used within internal communication can bring various communicative effects, particularly in relation to employee engagement.

In this thesis I will therefore analyze whether and how these communicative effects can be connected to internal communication. As such, this study aims to answer the

following research question:

How can stylistic features contribute to the communicative effect of employee engagement?

In answering this question, I largely adopt the research method of systematic style analysis as is used by Van Leeuwen in Stijl en Politiek. This analysis will be conducted on a corpus gathered from DSM. DSM (Dutch State Mines) is a multinational biotechnology company that has shown interest in optimizing its internal communication. In this thesis, the stylistic analysis focuses on texts taken from DSM’s internal communication hub: the internal DSM News Center. Through this News Center, DSM attempts to inform and involve its

employees in the happenings within the company. By focusing on the stylistic choices made in these news articles, the analysis provides insights into how stylistic choices help to support or enhance employee engagement in internal communication.

Next to the systematic style analysis, a reception analysis will be conducted with the help of a questionnaire. This questionnaire is aimed at creating indicative readers’

impressions of the News Center articles. With the readers impressions gathered, the stylistic analysis can evaluate which linguistic choices help to contribute to these impressions. The combination of these analyses will display how the communicative effects of linguistic features can potentially be connected to employee engagement, offering an intersection between the fields of stylistics and internal communication.

In chapter two, the theoretical framework, I will first delve into the goals internal communication attempts to achieve with its communication texts. Then, I will explore

(8)

4 Crystal & Davy’s (1969) concept of extra-linguistic contexts, and connect their three goals in achieving stylistic analysis to the News Center article analysis I am conducting. Finally, I will argue for the necessity of conducting a quantitative stylistic analysis with the help of a style checklist, and clarify the theory of Critical Discourse Studies, which will help guide the qualitative stylistic analysis.

In chapter three, I discuss the set-up of the reception analysis. First, I will cover the justification of the corpus, clarifying which news articles have been selected to be read by the questionnaire respondents. In the following paragraph, I provide an overview of recent academic insights in the connection between internal communication and employee engagement, and derive my criteria that will be tested in the questionnaire from this overview.

The process of the reception analysis will be discussed in chapter 4. I will discuss the results of the questionnaire, analyzing whether there are patterns and trends to be found in the answers provided by the respondents.

In chapter 5 I will discuss the process of the bottom-up and top-down phases of the style analysis. While the bottom-up phase uncovers all stylistic features that seem

particularly striking or apparent in the analyzed text, the top-down phase makes a selection out of these features, based on whether these features can be deemed relevant with regard to the respondents’ impressions brought by the questionnaire. In this chapter, I therefore outline which stylistic features are found bottom-up, and justify which of these features were selected during the top-down phase (and will therefore be covered in the analysis in chapter 6).

In chapter 6 I cover the systematic stylistic analysis of the corpus. The stylistic features selected during the top-down phase will be analyzed in the five corpus texts. Each feature will first be analyzed quantitatively per corpus text. Then, Critical Discourse Studies, together with academic insights gathered about these stylistic features, will guide the qualitative analysis, analyzing and interpreting the effect these stylistic features have in the corpus texts. Finally, chapter 6 outlines the results of the quantitative and qualitative analysis, uncovering whether any patterns or trends can be found.

(9)

5 In chapter 7, the results of the quantitative and qualitative analysis will be evaluated together with the reader’s impression provided by the questionnaire analysis. I will evaluate whether patterns can be found among the results, and to what extent these patterns display a potential relation between style and employee engagement. Chapter 7 will also discuss how the style analysis offers a stylistic point of view on the definition of internal communication as a extra-linguistic context. Finally, since this is an exploratory research, the considerable boundaries and limitations of this study will also be discussed.

First, I will outline the theoretical framework.

2. Theoretical framework

In this chapter, I will explore the theoretical framework of my thesis. First, I will gather recent insights on the definition of internal communication, outlining what goals internal communication follows to achieve employee engagement. Then, I will explore Crystal & Davy’s (1969) three goals in identifying linguistic features in a stylistic analysis. Finally, these goals will be connected to the field of Critical Linguistics, a part of Critical Discourse Analysis that focuses on the effects linguistic features can have on the social context they are used in.

2.1 Internal communication

Currently, the definition of internal communication is varied and defined by several

perspectives1_{. Among these perspectives, Vercic et al. (2012: 225) employed a closed-item} questionnaire, wherein internal communication is often equated with employee

communication. They state that respondents stressed that internal communication should consist of ‘an exchange in information among employees or members of an organization to create understanding […]’ (Vercic et al. 2012: 225). Moreover, respondents felt that internal

1 _{For an extensive summary on the current academic insights in internal communication, refer to Vercic et al.}

(10)

6 communication creates value for the organization by motivating employees, aligning the goals of individual employees to organizational goals. With regard to the practice of internal communication, respondents agreed that internal communication requires skills in data gathering and analyzing, oral presentations and speaking skills. The highest level of agreement was found in the idea that internal communication above all requires

communication skills such as designing, speaking and writing (Vercic et al.: 226). Welch and Jackson (2007: 494) define internal corporate communication as

‘communication between an organization’s strategic managers and its internal stakeholders, designed to promote commitment to the organization, a sense of belonging to it, awareness of its changing environment and understanding of its evolving aims’. If internal

communication can meet these goals of commitment, awareness and understanding, it can successfully contribute to employee engagement. Mbhele (2017: XIV) also stresses that internal communication is one of the key drivers of employee engagement. Welch (2012) focuses on an employee perspective of internal communication, and also stresses its relevance with regard to employee engagement:

Internal communication underpins organizational effectiveness since it contributes to positive internal relationships by enabling communication between senior managers and employees. Successful internal communication can promote employee awareness of opportunities and threats, and develop employee understanding of their organization’s changing priorities. (Welch 2012: 246).

Internal communication therefore succeeds in establishing a connection between

management and employees alike. Interestingly, internal communication can also become a source of irritation to employees. Welch (2012: 252) stresses that employees find a one-size-fits-all approach inappropriate, and as such internal communication must be catered to different workforces with different features, to employees with different education levels

(11)

7 and communication needs. Finally, Welch points out the fact that among other goals

‘persuasive writing and rhetorical messaging skills’ are a priority (253). This is in line with Vercic et al. (2012), who highlight that internal communication above all requires strong writing and speaking skills.

Interestingly, the ties between the goals of internal communication and the application of persuasive writing and rhetorical language skills has not been explored much, but the connection between internal communication and language use ‘is an issue that needs to be investigated and calls for strategic attention from organizations’ (Louhiala-Salminen & Kankaanranta 2012: 3). With the eye on globalization, research has been conducted on multilingual strategies within internal communication (Louhiala-Salminen & Kankaanranta 2012), questioning whether formal, ‘standard’ English should be employed in the organization’s intranet, when many readers might not be proficient enough to read formal English. Louhiala-Salminen & Kankaanranta do not elaborate on the connection between internal communication and persuasive writing.

Despite the lack of insights in the connection between persuasive writing and internal communication, these studies show that writing skills and language use are crucial aspects in establishing employee engagement through internal communication. As such, linguistic and stylistic features can both improve and deteriorate a company’s internal communication, but which features are likely to contribute to these factors has not been explored yet.

In the next paragraph, I will outline how a stylistic analysis, through a micro- and macro-level analysis can help gather insights on the potential effects that linguistic features can have within their social context.

(12)

8 2.2 Micro- and macro-level analysis

Style is often defined in what Anbeek & Verhagen call a macro-dimension (2001: 23), a general impression that readers have of a text when they are reading it. Anbeek & Verhagen (2001: 2) argue that authors often have their own stylistic signature, which can range from being unique and original to being trivial and clichéd. An example is a study conducted by Fagel, Van Leeuwen & Boogaart (2011), who analyze the macro-impression of Jan Arends’ ‘clear’ prose and Biesheuvel’s ‘chaotic’ style. Verhagen & Anbeek (2001: 23) stress that stylistic analysis is especially successful when it can demonstrate how the textual micro-level (such as words and grammatical constructions) produces impressions on a macro-level.

This is reinforced by Leech & Short (1981). They state the following in Style in

Fiction: ‘[…] by making ourselves explain how a particular effect or meaning is achieved

we not only understand better how it is achieved […], but also gain a greater appreciation of what the writer has created’ (1981: 2). Applied to the analysis of literary works, Leech & Short signify that stylistic analysis can offer us insights in how these macro-dimensions are created through linguistic and stylistic choices. A stylistic analysis therefore can help explain why an author’s writing is considered ‘clear’ or ‘chaotic’. Next to this, connecting an author’s writing to a ‘clear’ style will also grant us insights into which linguistic features are likely to have effects that contribute to the creation of this impression.

This concept of linguistic features on the micro-level contributing to the macro-level impression of a text can be connected to Crystal & Davy’s concept of an

extra-linguistic context: ‘everything non-extra-linguistic which exists at the time of using the extra-linguistic feature(s)’ (1969: 11). In Investigating English style, they cover these linguistic features, which occur at the micro-level of morphology, syntax, or composition and can help

contribute to the impression created in the extra-linguistic context. Different extra-linguistic contexts demand different stylistic choices, but it is up to the stylistician to explore why

(13)

9 these contexts demand these stylistic choices, and to analyze which effects these stylistic choices have when placed in that context.

To identify the effect of these linguistic features, Crystal and Davy (1969: 12) list three goals:

1. We must identify the entire range of linguistic features which people intuitively feel to be stylistically significant, and specify a precise way of talking about them (a

metalanguage).

2. We must outline a method of analysis which will allow us to organize these features in such a way as to facilitate comparison of any one use of language with any other 3. We must decide on these features, by classifying them into categories based on the kind

of extra-linguistic purpose they have.

Crystal & Davy define goal one as ‘identifying the entire range of linguistic features people feel to be stylistically relevant’. Any feature deemed relevant can be analyzed as a

contribution to the macro-level impression. I will further explore this goal in paragraph 2.2.1, by arguing that working systematically with the help of a style checklist ensures that linguistic features are identified on a consistent basis. The style checklist will also ensure that we can discuss the linguistic features in a precise and universal way.

Goal two focuses on outlining a method of analysis in such a way that it facilitates comparison. In this thesis, several internal news articles within the same context of internal communication will make up the corpus, thus facilitating working comparatively. The corpus, which will be further clarified upon in chapter 3, consists of texts that all cover a similar topic, and are thus easier to compare and contrast among each other.

Goal one and two are part of a quantitative research: they help us identify the frequency of linguistic features, and possible patterns between these uncovered features.

(14)

10 They do not help in interpreting the effects these features can have. In goal three, we

classify the effects these linguistic features have through a qualitative research: this will be explored through the study of Critical Linguistics, which is a part of the field of Critical Discourse Studies. In Paragraph 2.2.2 I will argue how Critical Discourse Studies provide valuable insights in assigning stylistic effects to linguistic features, by interpreting the results of the quantitative analysis as patterns contributing to the macro-level impression.

2.2.1 Quantitative research: style checklist and working comparatively

Crystal & Davy (1969: 12) define goal one as ‘identifying the entire range of linguistic features people feel to be stylistically relevant’. This goal can be translated to Van Leeuwen’s concept of a ‘bottom-up’ phase and a ‘top-down’ phase as part of the

quantitative stylistic analysis (2015: 36-38). In the bottom-up phase, the analyst does not take the macro-level impression of the texts into account yet – instead, the analyst only looks at which stylistic devices seem to be prevalent in the text. In the top-down phase, the macro-level impression is also taken into account. For each stylistic feature identified by the bottom-up analysis, the analyst will try to clarify which stylistic features might help contribute to the macro-level impression (Van Leeuwen 2015: 38).

Van Leeuwen stresses that the style analysis is done systematically (2015: 26). This can be achieved by using a style checklist. This list contains a selection of stylistic devices that have frequently lead to interesting insights in prior stylistic research (Leech & Short 1981: 61). The checklist therefore helps an analyst in ordering which stylistic devices in a text are relevant, based on for example relative frequency or position. The checklist, Leech & Short stress, is by no means exhaustive (1981: 61), but is a list of ‘good bets’, categories which based on prior research are likely to yield stylistically relevant information.

For the analysis in this master thesis, I will use the checklist that is set-up by Leech & Short (1981: 61-66). Leech & Short place the categories under four general headings:

(15)

11 A. Lexical categories

B. Grammatical categories C. Figures of speech, etc. D. Context and cohesion

The entire style checklist of Leech & Short can be found in appendix I. By using the checklist in the bottom-up phase of the analysis, the stylistician can go through each category and decide whether there are any categories that seem particularly striking in the analyzed text. For example, a text might contain a high amount of sentences in the passive form compared to similar texts, or these sentences might be placed at prominent positions in the text, such as in the opening and closing of the text (Van Leeuwen 2015: 37).

While the checklist helps in uncovering the frequency of different linguistic features, it is difficult to decide when a certain frequency is exceptional. Leech & Short cover this: ‘For example, discovering that x per cent of Gibbon’s nouns are concrete, and only y per cent abstract, is of little use by itself’(Leech & Short 1981: 41). Crystal and Davy also stress this through goal two of defining stylistic features within an extra-linguistic context: features must be organized in such a way that they can facilitate

comparison (1969:12). After all, working comparatively with more than one text will help in uncovering which stylistic features are exceptional when compared to other texts. Van Leeuwen also mentions this: for example, only by comparing text X to text Y, he stresses, can the analyst demonstrate that tekst X has a relatively big sum of gradable adjectives. (2015: 32-33) Since the corpus of this research is drawn from the same discourse (articles from the DSM News Center), linguistic features in certain texts can be more apparent especially because these features can be contrasted with other texts within the corpus – for example, text 5 might have a high amount of transitive verbs, whereas text 4 does not.

(16)

12 In the second part of the analysis, the top-down phase, the macro-level impression is taken into account. It is likely that the bottom-up phase of the analysis will bring up

stylistic features that cannot be deemed relevant to the established macro-level impression. In order to clarify whether a stylistic feature might be relevant to the macro-impression, the analyst often draws upon previous linguistic research, further exploring its potential

semantic and pragmatic attributes (Van Leeuwen 2015: 38). While the bottom-up analysis quantifies potentially relevant linguistic features, the top-down analysis attempts to uncover patterns in these features, elaborating how these features might show a stylistic pattern and how they might contribute to the macro-level impression. This also means that the top-down analysis is likely to narrow top-down the selection of stylistic devices uncovered through the bottom-up analysis (Van Leeuwen 2015: 52).

Van Leeuwen (2015: 27) signifies that while checklists help to give the analyst an overview of the stylistic occurrences in a text, they do not help the stylistician with interpreting the potential effects of these occurrences. Often the supposed effect of a stylistic device differs among contexts. Van Leeuwen exemplifies this by illustrating politician Geert Wilders’ strategy of naming his opponents’ first name as opposed to their last name – in many contexts this is a display of solidarity, but within political discourse it is rather a more belittling strategy. This means that linguistic features can have different effects in different contexts, and a checklist does not help in explaining what this effect can be. As such, a style checklist is not used for the interpretation of these stylistic features (Van Leeuwen 2015: 27).

In order to interpret the results of the quantitative research, to classify the effects of these linguistic features within their extra-linguistic context, I turn to two objectives. Firstly, insights in the potential effects of stylistic devices as concluded by earlier research will help guide the interpretation process. Examples of this include Lakoff & Johnson’s (2003) definition of conceptualizing metaphors and Fahnestock (2011) exemplifying the usage of various persuasive effects of stylistic devices such as the parallelism. These and more sources will be cited during the analysis itself in chapter 6.

(17)

13 Secondly, I cover the interpretative insights provided by the field of Critical

Linguistics, a part of Critical Discourse Studies, which will be covered in the next paragraph.

2.2.2 Qualitative research: Critical Discourse Studies

In line with goal three as established by Crystal & Davy (1969: 13), we must decide on the identified linguistic features, classifying which kind of effects these features have in the extra-linguistic context. The field of Critical Linguistics, which is a part of Critical

Discourse Studies (CDS), focuses on creating an awareness between linguistic aspects and their ideological effects (Wodak & Meyer 2016:6). Simpson (1993: 104), for example, covers the effect of transitivity in foregrounding certain meanings while covering up others by ascribing agency. CDS mostly employ these insights to analyze the major ideological effects that discursive practices may have, producing and reproducing unequal power relations such as men and women, cultural majorities and minorities (Wodak & Meyer 2016: 6). CDS therefore create awareness between stylistic choices and the effects they may bring on a macro-level context.

This last aspect of CDS will be used in the analysis in my thesis. In line with Van Leeuwen (2015: 18), the linguistic aspects uncovered in this research will not be analyzed for their ideological dimension. I will not address any political dimensions, but will adopt CDS’ notion that linguistic choices can bring about effects on the macro-level impression of a text.

CDS have frequently been criticized for being ‘ad hoc’ (Van Leeuwen 2015: 88), wherein an analyst often fails to argue why certain linguistic aspects are analyzed, and others are not. With the help of the quantitative analysis guided by a style checklist, the analyst can find stylistic features without excluding phenomena beforehand. Furthermore, by analyzing the stylistic features through working comparatively among the corpus texts, stylistic features are analyzed based on their relative quantities. These factors prevent the

(18)

14 analyst from ‘cherry picking’ linguistic features that happen to coincide with their own impressions.

On the level of a qualitative analysis, CDS teach us that analyzing stylistic choices on the micro-level of a text can help an analyst reach conclusions on the macro-level of a text (which is reinforced by Verhagen & Anbeek (2001: 23)). A common pitfall, however, is when an analyst lets prior ideas about the macro-level of a text influence the micro-level analysis. When these macro-level impressions are reached by a single person, the

objectivity of the research might be harmed. In other words, a stylistic analysis can be prone to bias: ‘The criticism rests primarily on what might be termed the interpretative

positivism shown by stylisticians who simply invoke linguistic description as a way of

confirming the decisions they have already taken about a text’s meaning’ (Simpson 1993: 111). To circumvent this, it is effective to establish a macro-level based on an

intersubjective dimension: this way the individual impressions of the stylistician will not influence the results of the micro-level analysis.

A good example of this is to conduct a reception analysis. A reception analysis is an effective, if extensive, way to achieve intersubjectivity (Van Leeuwen 2015: 26). By testing the corpus texts on a focus group through a questionnaire, we can establish an

intersubjective macro-level of the corpus with the results it gathers. In chapter 3 I will argue how interpretative positivism will be circumvented by conducting a reception analysis with the help of a questionnaire. The methodology behind this reception analysis will be

clarified in paragraph 3.1. and 3.2.

3. Method

For this thesis, the intersubjective dimension will be reached by conducting a reception analysis. The corpus is gathered from articles published on DSM’s internal News Center, which is effectively the nexus where all employees are informed of the newest happenings taking place in the company. The selection and justification of this corpus will be further

(19)

15 clarified in paragraph 3.1. These articles gathered from the DSM News Center will then be tested on an internal focus group: after reading each text, the respondents will rank the text on five different aspects of employee engagement. The criteria of these aspects can be found in paragraph 3.2.

Once the results have been gathered from the questionnaire, a macro-level readers’ impression will have been established. Then, the News Center articles can be analyzed quantitatively on stylistic choices that are likely contribute to this macro-level.

3.1 Corpus justification

For the questionnaire, five different texts varying in length from 229 to 553 words will be read by the respondents. These are all gathered from DSM’s Global News Center and can be found in appendix II. Five respondents will read these five texts and are then asked to rank these texts on employee engagement. One point of notice is that these texts are all different texts copied from the News Center, and not one text rewritten in five different styles. Therefore, the five corpus texts have differing content. As such, one potential danger is that respondents might rate these texts based on how compelling they find the content of the articles rather than the style or language that is employed.

To circumvent influence of the topic on the questionnaire results, all texts are selected to cover a similar topic, namely of DSM’s contributions to sustainability.

Sustainability is one of DSM’s key objectives and therefore a lot of articles written for the News Center cover this topic, which broadened the pool of eligible articles. To further prevent this topical bias, all articles cover news that occurred on sites that are far away (such as the opening of a solar field in Kingstree; none of the respondents live or have any direct affinity with the United States) or have been written more than a year ago (such as ‘June 2017: a month of sustainability achievements’). This might ensure that the content of the articles is less compelling to the reader, ideally increasing the relevance of language and style when the respondents rank the criteria on the questionnaire.

(20)

16 This way, I attempt to circumvent that the respondent ranks texts purely based on content. It must be stressed, however, that none of these criteria will clear up all topical influence: content will inevitably influence the perception of the reader, but with these criteria I attempt to foreground the stylistic effect over the topical effect. It is also important to note that the results of this questionnaire are used to create an indication, a reader’s impression of the news articles. This indication is used to help in guiding the quantitative and qualitative stylistic analysis, and to circumvent that I will conduct the stylistic analysis according to my own prior ideas on the articles. The questionnaire is employed to create an intersubjective dimension, to garner a global reader’s impression and to circumvent

interpretative positivism. The questionnaire, while necessary to create an intersubjective reader’s impression, therefore mainly serves to guide the analysis and evaluation conducted in chapter 6 and 7.

3.2 Questionnaire criteria

In this paragraph, I will discuss the criteria used to set up the questionnaire. While research on internal communication and employee engagement has been ongoing, little research on the relation between employee engagement and language use has been conducted.2_{As such,} this questionnaire approach is an exploratory research. Employee engagement in this thesis has been defined according to recent academic insights in internal communication, but is by no means definite. In this paragraph, I will now propose a definition of the goals of

employee engagement that is based on the goals DSM wants to achieve through internal communication. These goals will be supported and supplemented by recent academic insights.

2 _{For an overview on the current academic insights into internal communication and employee} engagement, refer to paragraph 2.1.

(21)

17 An interview I have conducted with Marianne Vintcent within DSM indicated that DSM uses internal communication to reach several goals. First, internal communication should always properly inform employees on the newest events and happenings within the company. Second, it is pivotal that internal communication helps in aligning employees with the brand and ideals of the business. Finally, it should help in instilling a sense of pride in employees: a DSM employee should feel encouraged to be working for the company.

These communication goals have been covered in prior research as well. Ruck & Welch (2011: 300), for example, created a conceptual model of employee questions in relation to employee engagement, although they do stress that this conceptual model is not yet definite (301). Additionally, not every aspect can be tested as Ruck & Welch (2011) often focus on the relation between employee and manager, and managers are not involved in this reception analysis. The model, however, conceptualizes engagement as an outcome of internal communication and ‘incorporates the importance of employee voice, based on being well informed, organizational support and identification’ (2011: 301). A selection of these aspects will be applied to the questionnaire in this thesis. A second study that focuses on the objectives of internal communication is the closed-item questionnaire of Vercic et al. (2012:225), where respondents were asked about their thoughts on the aspects that define the relationship between internal communication and employee engagement. Through these two researches I will further support the three criteria established through the interview with DSM, and supplement them with two more aspects gathered from these studies. This means that the questionnaire will cover five aspects in total.

DSM stresses that internal communication should always inform employees about the happenings within the company. Ruck and Welch (2011: 300-301) represent this in the first aspect of their model in the statement ‘I am well informed about what is going on and what is planned and my line manager is committed to the organization’. While the relation between employee and manager could not be tested in this questionnaire, testing how well informed a reader feels after reading a News Center article can provide valuable insights with regard to the effect of internal communication on employee engagement. This is also

(22)

18 supported in the reception analysis conducted by Vercic et al. (2012: 225), where

respondents answered that internal communication plays a pivotal role in the dissemination of information. These goals align with DSM’s objective of keeping employees

well-informed. Therefore, the questionnaire respondents will be asked to answer whether they feel well informed after reading the News Center articles.

Vercic et al. (2012:225) concluded that respondents argue for the importance of internal communication in aligning employees with the organization’s purpose. This ties into DSM’s objective of using internal communication to align employees with the company’s brand and ideals. This aspect has been supplemented with a finding in the research of Ruck & Welch (2011: 300-301). While DSM and Vercic et al. (2012) suggest the importance of alignment, Ruck & Welch (2011) also state that employee identification is a pivotal aspect of internal communication, translated through the statement ‘I identify with the organization’s values and am an advocate of what it does’. In the questionnaire used for this thesis, both aspects will be tested: respondents will be asked to rank whether the corpus texts allow them to identify with the values of the company, and will also be asked to rank whether the text helps them align with the company’s values.

DSM uses internal communication to ‘instill a sense of pride’ in its employees. Neither Vercic et al. (2012) nor Ruck & Welch (2011) state that ‘instilling pride’ is a goal of internal communication. Instead, Vercic et al. (2012: 225) stress that respondents connected the objective of internal communication to ‘motivating employees’ as a way of creating value for the company. Since both definitions are distinct, I decided to include both aspects in the questionnaire. The goal of ‘motivating employees’ is directly derived from Vercic et al. (2012), whereas testing ‘instilling pride’ is used to satisfy DSM’s objective. This means that both these goals of ‘motivating employees’ and ‘instilling pride’ will be tested in the questionnaire.

In total, five aspects will be tested in the questionnaire. Three aspects have been gathered from DSM and have been supported by the insights provided by Ruck & Welch

(23)

19 (2011) and Vercic et al. (2012): information dissemination, employee alignment, and

employee pride. Furthermore, these criteria have been supplemented with two more aspects provided by these studies: employee identification and employee motivation. The five aspects can be found below:

1. Information dissemination 2. Identification

3. Alignment 4. Motivation 5. Pride

These aspects have been translated into the following criteria in the questionnaire:

1. I feel well-informed after reading this text.

2. Reading this text helps me identify with DSM’s values as an organization. 3. Reading this text helps me align with DSM’s purpose.

4. This text motivates me as an employee working for DSM 5. Reading this text makes me feel proud to be a part of DSM.

In the following chapter, I will elaborate how these criteria have been tested in the questionnaire. Additionally, I will briefly discuss whether the results of the reception analysis indicate any general patterns or trends in the questionnaire scores.

(24)

20 4. Questionnaire analysis

During the questionnaire, the respondents were asked to rank the five corpus texts on five aspects through a 7 point Likert Scale, with 1 signifying that the respondent does not agree with the aspect at all, while 7 means the respondent identifies very strongly with the aspect. In this chapter, I will elaborate on the reception process, and cover the results that spurred from this questionnaire analysis. Then, I will discuss whether there are trends or patterns that can be derived from the questionnaire results, and argue how they can provide an indicative macro-level reader impression that will help guide the systematic stylistic analysis and evaluation in chapter 6 and 7.

4.1 Description of the reception process

With the help of DSM internal communications officer Marianne Vintcent I received a list of seven email addresses from employees within DSM. These employees are all part of DSM NEXT, a global network of over a thousand young professionals in DSM. DSM requested that the questionnaire was answered by respondents representing DSM NEXT, largely because they are of a young age (21-31).

The employees were approached to fill in the questionnaire through email. Out of the seven employees, five (3M/2F) answered the questionnaire. Since DSM NEXT is a global network, the employees live in five different countries: Brazil, France, Italy, India and Switzerland. These respondents answered the questionnaire anonymously. Respondents were asked to read the text, and then immediately thereafter had to rate the text on the Likert scale based on the employee engagement aspects as discussed in paragraph 3.2. This happened across all five texts: the respondents were therefore requested to fill in the aspects of employee engagement five times in total. In the next paragraph, I will cover the results of internal testing.

(25)

21 4.2 Results of internal testing

The results of the questionnaire analysis have been visualized into five different graphs per text: these represent the five different aspects of employee engagement as discussed in chapter 3.2. The graphs can be found in appendix III. By ordering the different aspects in separate graphs per text, we can see more clearly the common trends between the texts, signifying which texts scored better in the questionnaire and which scored worse.

One point of notice is that, with the exception of figure 5.2 (appendix III), all aspects rated as 4 or higher on the Likert scale, meaning there was almost no negative feedback given on any of the five texts. A potential reason for this occurrence could be that although the questionnaire was stressed to be anonymous, employees were approached individually through email. As such, employees might have been reluctant to provide negative feedback. I will cover this pitfall in greater detail in chapter 7.

While the respondents did not provide negative feedback, variations between the scores of 4 and 7 can be spotted among the five different texts. For example, text 2, ‘Grand Opening of the Kingstree Solar Field!’ almost exclusively scores a 6 or 7 on the Likert scale, whereas text 5, ‘Accountability for sustainability’ has a more varied response: scores range from 4 to 7, and as signified in figure 5.2, text 5 is the only text to score a 3 on employee identification.

It must be stressed that the results from the questionnaire research are merely an indication: they grant us a rough idea about which texts were preferred by the five respondents. The pool of respondents, however, remains very small, meaning sweeping statements about the results of this questionnaire cannot be made. The questionnaire, however, does help us in giving an indication of an intersubjective macro-level impression: despite the low sample pool, text 2 and 4 are rated higher on employee engagement, and this impression is established through an intersubjective opinion, meaning the questionnaire is nonetheless helpful in circumventing interpretative positivism.

(26)

22 Since the sample size is low and the definitions of the goals of employee

engagements are proposed and not yet set in stone, it is complex to derive conclusions of the questionnaire scores of individual questions. Instead, a general pattern in the

questionnaire indication has been uncovered by rating the texts not on their individual questions, but on their total scores. The scores visualized in the table below are calculated by adding the scores of each question together, then dividing them by the total amount of ratings. To illustrate, question 1 of text 1 has a total score of 29 (4+5+6+7+7). This is applied to all five questions of the text, then divided by the total amount of ratings (25) to reveal an average score: 144 (average 5.76). In table 1, the left column displays the five different corpus texts. The right column displays the total and average scores of the

questionnaire rating of each text. Additionally, the texts are ordered based on their rating on employee engagement:

Table 1. Results of the reception analysis

Article Total score (average)

Text 2. Grand Opening of the Kingstree solar field!

162 (6.48)

Text 4. June 2017: A month of sustainable actions

161 (6.44)

Text 3. A Sustainability success story 148 (5.92) Text 1. #beatplasticpollution for World

Environment Day

144 (5.76)

Text 5. Accountability for sustainability 131 (5.24)

As can be derived from table 1, text 2 and 4 are the highest scoring texts with an average of 6.48 and 6.44 respectively, while text 5 is the lowest scoring text at 5.24. Text 1 and text 3 score 5.76 and 5.92 respectively. The questionnaire therefore indicates that text 2 and 4

(27)

23 were rated the most positively with regard to employee engagement, whereas text 5 was rated the lowest.

This indication provided by the questionnaire establishes an intersubjective macro-level reader’s impression. With this impression established, it is now possible to conduct a style analysis. Through this analysis, we can uncover potential patterns between the texts that score higher or lower in employee engagement and the stylistic features that are

employed in these articles. This style analysis will be conducted in chapter 6. Then, we will analyze in chapter 7 whether a trend can be found between the macro-impression derived from this questionnaire, and the stylistic features uncovered in the five corpus texts. In the following chapter, I will discuss and justify the selection of stylistic features that will be analyzed in the style analysis.

5. Bottom-up and top-down analysis justification

In this chapter, the selection of stylistic features in the process of the bottom-up and top-down analysis will be justified.

5.1 Selection in the bottom-up analysis

As discussed in paragraph 2.2.1, the stylistic analysis has been conducted systematically with the help of a checklist, through a bottom-up and a top-down analysis. In the bottom-up phase, I have looked at all the stylistic devices that were at a first glance prevalent in the text (based on their quantity or position in the text). During the top-down phase, all these features were evaluated based on two criteria: whether their communicative effects (established through prior academic research) could potentially be linked to the five goals of employee engagement, and whether the corpus texts showed a variation in the

(28)

24 clarify and justify how the stylistic features covered in the bottom-up analysis were

narrowed down from ten to four features in the top-down analysis.

In table 2 found below, all linguistic features covered during the bottom-up analysis have been gathered in the left column. The right column displays the relation between the stylistic feature and its category on the style checklist:

Table 2. Stylistic features analyzed during the bottom-up analysis

Linguistic feature Category on the style checklist

Concrete and abstract nouns Lexical categories: A2 Frequency and attributes of adjectives Lexical categories: A3 Stative and dynamic verbs Lexical categories: A4

Transitive verbs Lexical categories: A4

General types of grammatical constructions Grammatical categories: B9 Grammatical and lexical devices Figures of speech: C1

Phonological schemes Figures of speech: C2

Tropes (metaphor and personification) Figures of speech: C3

Avoidance of repetition Context and cohesion: D1

Context (attitude of the author) Context and cohesion: D2

What immediately stands out in table 2 is that the bottom-up phase largely uncovered lexical features on a word-level: four categories of lexical features were analyzed, along

(29)

25 with three categories of figures of speech, which within this corpus often occur on word-level rather than clause-word-level. Finally, category D1, avoidance of repetition, covers the usage of synonyms and similar descriptive phrases and thus also largely functions on word-level rather than clause-word-level. This absence of grammatical categories will be further discussed during the evaluation in chapter 7.

Four stylistic features have been selected during the top-down analysis: concrete and abstract nouns, transitive verbs, tropes and avoidance of repetition. These stylistic features will be explored on their communicative effects and their potential connection to employee engagement in paragraph 5.2.1 to 5.2.5. First, I will argue why the other six categories of stylistic features were not selected during the top-down analysis.

Starting with lexical category A3, frequency and attributes of adjectives, all

adjectives were counted and furthermore categorized between restrictive and nonrestrictive adjectives, gradable and nongradable attributes, and attributive and predicative positions. Each corpus text, however, scored similarly on the different categories: the overwhelming majority of adjectives in each text was restrictive, nongradable and attributive. This means that only the category of ‘adjectives’ as a whole could be quantified. Van Leeuwen (2015: 34) states that general linguistic categories such as ‘questions’, ‘nouns’ or ‘imperatives’ are abstract and uninformative: quantifying these general categories does not help in

uncovering their semantic function. This applies to the adjectives in the corpus texts as well: they could not be distinguished based on their semantic function. Therefore, the corpus did not facilitate working comparatively.

While the category of stative and dynamic verbs did show variation (text 1 and 5 contained a relatively high amount of stative verbs as opposed to dynamic verbs), I questioned whether this category would overlap too much with the category of transitive verbs. Both categories display a similar trend, with text 1 and 5 being more nominalized, and less driven by transitive verbs while text 2 and 4 contained a higher amount of dynamic and transitive verbs. Furthermore, the category of verb transitivity has been studied

(30)

26 extensively within Critical Linguistics and Critical Discourse Studies for its foregrounding, agency ascribing communicative effects (Fowler & Kress 1979: 198, Simpson 1993: 87), providing a potentially useful communicative effect in relation to employee engagement (this will be further motivated in paragraph 5.2.2). Therefore, I opted to choose for the category of transitive verbs to be covered during the top-down analysis, favoring it above the category of stative and dynamic verbs.

The category of grammatical constructions displayed potentially meaningful results: text 3, for example, contained a considerable amount of comparative constructions and text 4 often used enumerative structures and text 2 frequently used constructions of causality. It was difficult, however, to compare the corpus texts within this category, since each text contained a different majority of grammatical constructions (either comparative, causal, or enumerative). Furthermore, while I speak of a ‘majority’ of grammatical constructions, this ‘majority’ was often no more than one or two constructions per text. Deriving systematic results from the quantitative analysis therefore was complicated and arbitrary. It is for this reason that the category was dropped during the top-down selection.

Out of category C of the style checklist (figures of speech), only the category of tropes was used in the majority of the texts. Of the grammatical and lexical devices, it was rarely possible to find a stylistic device that was used in more than one text. For example, one parallelism was found across the five corpus texts (text 2: ‘fulfilling our commitment […] upholding our responsibility’), making it impossible to work comparatively.

Initially, the tricolon appeared to be a potentially interesting stylistic feature because it suggests closure or completeness of a topic (Charteris-Black 2014: 41) and this feature was found in most corpus texts (especially in text 2, 3 and 4). After taking a closer look at these tricolons, however, it begged the question whether these stylistic phenomena could truly be considered as tricolons. For example, ‘[…] electricity including solar, wind and hydroelectric’ (Text 2, 18-19) was initially considered as a tricolon, but upon a closer look

(31)

27 appears to simply be an enumeration of three informative aspects rather than a combination of words used to make a strong rhetorical impression.

Some stylistic features were found in category C2 of phonological schemes as well, such as alliteration and assonance. Many occurrences often seemed to be accidental rather than a deliberate stylistic choice: while examples such as ‘peak production’ and ‘renewable resources’ can be read as alliterations, they were used sparingly and it is debatable whether infrequent instances of alliteration can have a comparable communicative impact to for example use of metaphors (a category that has been selected in the top-down analysis, under ‘tropes’).

Finally, category D2 of the style checklist seemed potentially interesting for the way the reader is addressed by the writer. This was mostly guided by the occurrence in text 3: here, two employees are frequently directly quoted, and as such the text does not use an ‘inclusive we’, but instead makes use of the third person. The other four texts, however, all universally use the ‘inclusive we’, making text 3 an exception. This was a category of interest since use of the feature of ‘inclusive we’ has been analyzed for its potential to reduce the deniability of otherwise refutable assertions, reinforcing the effect of evaluative statements (Mulderrig 2011). The nature of the text (two employees from a specific site explaining their sustainability measures), however, already clarifies why the third person was used as opposed to an ‘inclusive we’. Since text 3 was the only exception, it was difficult to establish working comparatively among the texts: the other 4 texts show no variance in the addresser-addressee relationship and universally use an inclusive-we. Had the variance among the texts been greater, this category could have been further analyzed during the top-down phase.

(32)

28 5.2 Selection in the top-down analysis

During the top-down analysis, the stylistic features covered in paragraph 5.1 were filtered out. This means the top-down analysis resulted in four stylistic features. These features, together with their connection to the style checklist, can be found in table 3 below:

Table 3. Linguistic features selected during the top-down phase

Linguistic feature Category on the style checklist

Concrete and abstract nouns Lexical categories: A2

Transitive verbs Lexical categories: A4

Tropes: metaphor and personification Figures of speech: C3 Avoidance of repetition Context and cohesion: D1

The linguistic features found in table 3 have been selected based on two criteria. In paragraph 5.2.1 to 5.2.4, I will argue that these features have established communicative effects that can potentially be linked to the aspects of employee engagement discussed in paragraph 3.2. Furthermore, as will become evident in chapter 6 (analysis), the corpus texts showed variance in the usage and application of these linguistic features, thus facilitating working comparatively.

Since the relation between employee engagement and stylistics has rarely been explored, it is difficult to decide how uncovered stylistic patterns might play a role in the communicative effect of internal communication. We might for example find that text 2 contains a high amount of concrete nouns and transitive verbs, whereas text 5 has a high degree of abstract nouns and nominalization, but no study has been conducted on the potential relation between nominalization and internal communication or employee

(33)

29 engagement. Research has shown, however, that nominalization often contributes to the ‘vagueness’ of a text (Onrust 2013: 19-20), which might conflict with the employee engagement aspect of information dissemination. It must be stressed that the potential stylistic effects identified in the top-down analysis are discussed in this thesis as a first bridge or indication of the potential relation between stylistics and employee engagement. Therefore, this topic is certainly subject to future research, and will be further evaluated in chapter 7.

Van Leeuwen (2015: 46) states that the style analysis of political language has often been focused on common lexical features and stylistic devices, such as transitivity,

nominalization, and use of the metaphor, while he displays that stylistic features on a grammatical level are also worth exploring. In the case of internal communication, a connection between this field and stylistics has rarely been established. Therefore, I

deliberately focus on these ‘common categories’. As academic consensus on these stylistic features is well-established, they can function as a strong stylistic foundation in uncovering potential connections and patterns between stylistic features and employee engagement.

In order to uncover the connection between stylistic features and the goals of employee engagement, I will now take the stylistic features as outlined in the previous paragraph, and connect them to the goals of employee engagement covered in paragraph 3.2. As discussed in paragraph 3.2, the following aspects of employee engagement were derived: 1. Information dissemination 2. Identification 3. Alignment 4. Motivation 5. Pride

(34)

30 In the following paragraphs, I will now cover the academically established communicative effect of each stylistic feature, and connect them to one or several goals of employee engagement. After these connections have been covered in paragraph 5.2.1 to 5.2.5, I will then analyze how these stylistic features are used across the corpus texts in chapter 6.

5.2.1 Concrete and abstract nouns

The first stylistic feature, concrete and abstract nouns, can play an important role in

conveying a clear and coherent story. Burger & De Jong (2009: 44) stress the importance of a ‘show, don’t tell’ approach when writing a text. This can be achieved by using concrete, specific and clear language use (2009:44). They stress that this is especially important for genres such as columns, reports and travel stories, because these genres depend especially on helping the reader to envision for himself what is being written (44 ) – a landscape might be described as ‘beautiful’, but only by describing it’s snow-capped hills or lush green trees can the reader imagine why the landscape is beautiful. Similarly, as discussed in paragraph 3.2, internal communication relies on informing employees about the happenings within the company across the globe. As such, similar to the genre of a news report, the news articles rely on conveying a story that can help the reader imagine the scenario as if he was present himself. Only by involving the employees into the topic at hand will they feel engaged about the events happening at DSM.

One way of conveying a concrete story is by distinguishing between concrete and abstract nouns. Where concrete nouns have a direct relation with names of people, animals, things or materials (E-ANS, 3.2.1), abstract nouns refer to entities that cannot be observed through the five senses, such as ideas, concepts and feelings (E-ANS, 3.2.1). Onrust (2013: 18) covers how abstract nouns do not have a direct relation to the concrete reality, which can threaten the degree to which a text manages to help the reader envision what the text is really about. Abstract nouns can therefore cause a text to become vague, and threaten the

(35)

31 ‘show, don’t tell’ approach. On the flip side, concrete nouns can help a reader envision what is happening in the story.

Since concrete nouns can therefore create a concrete image for the reader to envision, I argue that concrete nouns can support the first goal of employee engagement: information dissemination. Abstract nouns can work against this goal by causing a text to become more vague.

5.2.2 Verb transitivity

As part of the style checklist category A4 (verbs), Leech and Short (1981: 61) focus on whether verbs are transitive or intransitive. Verb transitivity has been studied extensively within Critical Discourse Studies (Fowler & Kress 1979: 198, Simpson 1993: 87) and has been applied to the analysis of news articles as well (Teo 2000: 25). The corpus analyzed in this thesis, which covers the context of news articles and news reports within internal communication, is therefore a similar body of text in which the mechanics of transitivity can be analyzed.

Simpson clarifies how transitivity ‘has been employed to uncover how certain meanings are foregrounded while others are suppressed or obfuscated’ (1993:104):

transitivity can determine the degree of involvement the subject has in the action applied to the object. For example, while ‘I broke the vase’ ascribes agency to myself and therefore makes me the explicit culprit, changing the sentence to ‘The vase was broken’ leaves my involvement and agency implicit (1993: 94). This way, using transitivity helps in ascribing action and creating agency. Teo employs this technique in uncovering how news articles consistently ascribe action to police officers, whereas the supposed criminals are described through passivization, removing agency and therefore diminishing their active role in the article (Teo 2000: 27).

(36)

32 While Teo employs a transitivity analysis to uncover racism in the news, I will apply this notion of transitivity in relation to the goals of employee engagement. Since transitivity helps in ascribing agency to subjects, transitivity can potentially contribute to employee engagement by helping to keep a text more engaging. Through transitivity, the ‘doer’ of the sentence is made explicit, thus foregrounding the role of the actor. For DSM’s internal news articles, this means that transitivity can effectively foreground the role of DSM as an actor, elaborating the goals DSM has achieved.

As such, verb transitivity could satisfy two employee engagement goals: since DSM, the common actor in the news articles, is foregrounded, it can help the reader to identify with the company’s values (goal two). Furthermore, since texts with a high degree of transitivity will explicitly connect achievements to the company (the actor), transitivity could also instill pride in an employee reading the news article (goal five).

5.2.3 Tropes: metaphor and personification

Leech and Short (1981: 63-64) cover category C on the style checklist as ‘figures of

speech’. This category ranges from grammatical and lexical schemes to tropes. Tropes draw upon the ‘associative power of language’ by turning words away from their expected

definition and instead drawing attention to other potential meanings (Charteris-Black 2014: 45). A trope that has been covered extensively in academic literature is the metaphor. Imagery such as the metaphor can help in breaking down abstract ideas into concrete explanations (Burger & De Jong 2009: 75). Furthermore, the metaphor is covered as ‘the prime rhetorical figure for evoking emotions appropriate for the political purpose of

motivating social action’ (Charteris-Black 2014: 49): metaphors can therefore play a crucial role in enabling social and political responsibility.

Lakoff & Johnson (1980: 156) state that metaphors are adept at creating social realities, meaning a metaphor ‘may thus be a guide for future action’. They exemplify this with President Carter’s coverage of the energy crisis, phrasing it as ‘the moral equivalent of

(37)

33 war’. By conceptualizing the energy crisis as a war, the metaphor creates several

implications (or ‘entailments’ (Lakoff & Johnson 1980: 156)), for a war requires an enemy, and enables a need for resources and creates the narrative of a fight for survival. This mechanism of metaphors inspiring political and social action will be further explored with regard to employee engagement.

In paragraph 6.2.3, two types of imagery will be uncovered and analyzed in the News Center articles of DSM. Firstly, DSM uses a metaphor in representing sustainable development as a ‘journey’ and a ‘commitment’. Secondly, when discussing the importance of sustainability, DSM uses personification (ascribing human qualities to non-human entities (Charteris-Black 2014: 47)) to represent the environment as an entity that can be ‘healed’.

I will argue in paragraph 6.2.3 that both stylistic features are employed to inspire social action, to appeal to an individual responsibility of the reader. In doing so, the imagery used can satisfy two goals of employee engagement. It can stimulate employee alignment (goal 3), because representing sustainable development as a ‘journey’ means that the process is ongoing and has not reached its destination yet, engaging readers to keep working hard for a common cause. Furthermore, it can encourage employee motivation (goal 4) by appealing to the reader’s sense of individual responsibility: the personification of the environment helps to represent the environment as a concrete entity, ensuring ‘sustainability’ is no longer an abstract phenomenon, but a concrete process that the reader can participate in.

5.2.4 Avoidance of repetition

Under cohesion (D1) Leech and Short (1981: 64) raise the following point: ‘Alternatively, is any use made of elegance variation – the avoidance of repetition by the substitution of a descriptive phrase […] Are meaning connections reinforced by the repetition of words and phrases or by repeatedly using words from the same semantic field?’ In style and writing

(38)

34 guides it is often discouraged to repeat the same word (Burger & De Jong 2009: 171,

Renkema 2012: 131). Moreover, using synonyms can efficiently provide the reader with additional information about the text. (Burger & De Jong 2009: 171). Leech and Short also name how repetition can be avoided by substituting a word with a descriptive phrase (1981: 64). They exemplify the use of a descriptive phrase by mentioning how ‘her uncle’ (who is a lawyer) can later be substituted for ‘the old lawyer’, providing new information to the reader without extending sentence or word count.

Avoiding the repetition of words can therefore be a stylistic strategy with a beneficial communicative effect for employee engagement. Out of the five aspects of employee engagement, avoidance of repetition can support the objective of information dissemination: by using synonyms or substitutions a text can more efficiently inform the reader. While applied to texts of the literary genre, Leech and Short (1981: 85) stress that elegant variations are ‘not merely long-winded substitutes for a name: they draw attention now to this, now to that, aspect of the same person, and so build up a many-sided picture of each character’. Here, I argue that internal communication texts containing this ‘elegant variation’ can bring about a similar stylistic effect: by substituting words with synonyms or similar descriptive phrases, an internal communication text can stress multiple aspects of an event without increasing the text’s word count. This can help to make a text more

informative, reinforcing goal 1 of employee engagement.

5.2.5 Stylistic features and the goals of employee engagement

As discussed in paragraph 5.2.1 to 5.2.4, the four stylistic features have been connected to the five goals of employee engagement. The potential connection between these features and goals can be found in table 4 below:

From word to workplace. An exploratory research on language style and employee engagement.

From word to workplace

An exploratory research on language style and employee engagement

Table of contents