Building a Trustworthy Corporate Identity: A Corpus-Based Analysis of Stance in Annual and Corporate Social Responsibility Reports

Building a Trustworthy Corporate Identity: A Corpus-Based Analysis of Stance in Annual and... Abstract This article presents a corpus-based analysis of stance in annual and corporate social responsibility (CSR) reports. It investigates how companies use stance expressions to construct and promote a positive corporate identity to gain the trust of the stakeholder groups that these texts target. The results show that companies profile distinct identities in annual and CSR reports. In annual reports, they use stance resources to portray themselves as unbiased, rational, and competent decision makers. In CSR reports, they present themselves as committed, honest, and caring corporate citizens. These discursive identities are interpreted as strategic self-representations that optimize the persuasive appeal of the reports by addressing the specific expectations of the target readerships. This study sheds some new light on the identity work performed by companies in their public discourse. It also provides novel insights into the impression management strategies used by companies in annual and CSR reports. Finally, it provides both linguists and business communication scholars with a robust descriptive basis for critically assessing financial and CSR reporting. 1 INTRODUCTION Companies use communication strategically to control and positively influence the image the public has of them, and build favorable relationships with the stakeholders on which they depend (Van Riel 1995; Hallahan et al. 2007; Merkl-Davies and Brennan 2007). One key strategy they use to pursue these communicative goals is to project a positive corporate identity (Van Riel 1995; Fombrun 1996). Corporate identity refers to ‘a firm’s strategically planned and purposeful presentation of itself’ (Westcott Alessandri 2001: 177). By communicating a positive corporate identity, business organizations seek to create a positive image in the minds of the public and, over time, develop a good reputation (Van Riel 1995; Fombrun 1996; Westcott Alessandri 2001). A good reputation, in turn, can provide numerous benefits to organizations, including increased competitiveness and profitability, and a greater ability to attract customers, investors, and talented employees (for a review, see Walker 2010). Companies have multiple stakeholders with different, and often competing, expectations and demands (Freeman 1984; Friedman and Miles 2006). Investors, for example, pursue profit maximization and risk minimization; employees primarily seek identification, safety, and job satisfaction; customers demand product quality and distinctiveness (Johansen and Ellerup Nielsen 2011). But different stakeholders also have different expectations as to how a company should be in terms of its characteristics and values; that is, they have specific expectations concerning a company’s corporate identity. For instance, investors attribute considerable importance to managerial competence, employees value transparency and benevolence, and customers prioritize technical competence and integrity (Pirson and Malhotra 2011). Companies need to understand and fulfill their key stakeholders’ expectations to gain access to vital resources and be successful (Dowling and Pfeffer 1975; Freeman 1984). This article explores how companies use discourse to manage their stakeholders’ expectations and build trust with them. More specifically, it investigates how they use stance expressions (Biber and Finegan 1989; Biber et al. 1999; Conrad and Biber 2000) to discursively construct a trustworthy corporate identity in their communications with different stakeholder groups. The term stance refers to the linguistic encoding of personal feelings, attitudes, and evaluations (Biber et al. 1999). Stance resources are directly implicated in how speakers1 communicate their identity and negotiate interpersonal relations (Thompson and Hunston 2000; Martin and White 2005; Englebretson 2007; Bednarek 2015), and are therefore assumed to play an important part in the discursive construction of corporate identity. The study focuses on two major genres of corporate public discourse, that is, the annual report and the corporate social responsibility (CSR) report. The former is primarily addressed to shareholders and investors, and includes information about a company’s financial standing. The latter is targeted at a broader and more heterogeneous audience of stakeholders, and provides information about a company’s social and environmental performance. The analysis compares the use of a broad range of stance devices in a specialized corpus of annual and CSR reports recently published by a sample of large multinational corporations. It addresses the following research questions: Q1: Are there differences in the frequency and type of stance expressions used in annual and CSR reports? Q2: How are stance expressions used to discursively construct and promote a trustworthy corporate identity in these two text types? Q3: Do companies present themselves differently in annual and CSR reports? And, if so, how do they discursively shape their identity to appeal to and earn the trust of the audiences that these texts target? The article offers several contributions to the literature on corporate identity and corporate communication. First, it extends previous discourse analysis work on corporate identity (Koller 2009; Fuoli 2012; Bondi 2016), by (i) examining the role of stance in the discursive process of corporate identity building, and (ii) investigating how companies strategically shape their discursive selves to meet the expectations of different stakeholders. Second, it provides novel, linguistically informed insights into the impression management strategies deployed by companies in their reports (Neu et al. 1998; Hooghiemstra 2000; Merkl-Davies and Brennan 2007). Third, the study extends our knowledge of the genres of annual and CSR reports by providing a systematic descriptive account of stance in these text types, and by identifying differences between them. Finally, the study contributes to the critical literature on corporate reporting (Livesey and Kearins 2002; Lischinsky 2011; Deegan 2014; Jaworska and Nanda 2016), by exposing some important persuasive—and potentially manipulative—discursive tactics used in annual and CSR reports. The article is organized as follows. Sections 2 and 3 review previous discourse analysis work on annual and CSR reports, respectively. Section 4 introduces the concept of stance and outlines the analysis framework. Section 5 describes the corpus. Section 6 discusses the method used to identify and quantify stance expressions. Section 7 presents the results of the analysis. Finally, section 8 discusses the main findings and section 9 offers conclusions. 2 ANNUAL REPORTS The annual report is a formal document that incorporates both legally required and voluntary information about a company’s financial situation and future prospects. It is explicitly addressed to the company’s shareholders, although the readership is larger, and includes public and private investors, financial experts, government authorities, specialized media, brokers, creditors, and employees (Anderson and Imperia 1992; Ditlevsen 2012b). The annual report is generally regarded as a ‘hybrid’ discourse genre that combines informative and persuasive elements (Bhatia 2004; Garzone 2004; Rutherford 2005; Malavasi 2010; Breeze 2013). It can be seen to fulfill two main communicative purposes: (i) to give a true and fair view of the state of the company’s affairs, and (ii) to provide a positive image of the company (Ditlevsen 2012b: 97). The different sections within the report play different roles in this respect. Some sections have a primarily informative function, while others are mainly devoted to promoting the company’s image (Breeze 2013: 87). In general, narrative sections containing discretionary information tend to be more overtly promotional. For example, the CEO letter, which normally introduces the report by providing an overview of the company’s performance, is typically explicitly rhetorical and optimistic (Hyland 1998; Garzone 2004, 2005; Breeze 2012; Fuoli and Paradis 2014). Conversely, sections that contain mandatory or audited financial statements have a more neutral, objective, and technical tone (Breeze 2013: 87–8). While the annual report remains ‘first and foremost a statutory document’ (Ditlevsen 2012b: 110) that is used by companies to meet their legal requirement to inform shareholders about their financial situation, diachronic research has shown that its promotional function has gained increasing prominence (Beattie et al. 2008; Ditlevsen 2012b). The annual report is now commonly used by companies as a public relations tool to build a positive corporate image and promote trust (Rutherford 2005; Beattie et al. 2008; Malavasi 2010; Ditlevsen 2012b; Breeze 2013). This transition is evidenced by the increasing number of pages, texts, and pictures included in the annual report (Beattie et al. 2008; Ditlevsen 2012b). As Beattie et al. (2008: 186) observe, the annual report has changed over time from ‘a financially driven, statutory document toward a more design-oriented document’. It has also become increasingly multimodal (David 2001; Garzone 2004; Rutherford 2005; Ditlevsen 2012a,b), and it is today a ‘complex and colorful publication’ (Breeze 2013: 85). Given the important role the report has in orienting investors’ decisions, this shift is not uncontroversial. The informative and promotional goals of the annual report can sometimes conflict with one another. Graphs and pictures, for example, have been found to be often used as impression management devices, that is, ‘to convey a more favorable impression of the organization than is warranted’ (Beattie et al. 2008: 183). Therefore, the overtly persuasive nature of the modern annual report has often been criticized (Anderson and Imperia 1992; Breeze 2013). The promotional function of annual reports is complex and multifaceted. First and foremost, through the annual report, the company wants to present itself as an attractive investment to the shareholders and potential investors, encouraging them to maintain or increase their financial support (Ditlevsen 2012b; Breeze 2013). As Ditlevsen (2012a: 379) puts it, in the annual report ‘a company displays its successes and the investment potential of its shares in order to give an impression of its ability to succeed in the future’. More broadly, the annual report is used as a vehicle for constructing and promoting a positive corporate identity (Hyland 1998; Malavasi 2010; Ditlevsen 2012a; Breeze 2013). Companies discursively ‘engineer’ a positive corporate self by emphasizing their positive attributes, aims, and values (Breeze 2013). In addition, the annual report has a relational and rapport-building function; it is used to foster good relationships with investors and other stakeholders, to encourage confidence in the company’s management team, and to build trust in the company (Malavasi 2010: 212; Breeze 2013: 84). The language used in annual reports combines purely referential elements with affective and evaluative elements. The latter tend to be more prominent in the non-technical narrative sections (Garzone 2004). Positively charged words and expressions are frequent, and are used to highlight the company’s achievements, competitiveness, and ethical values (Malavasi 2010). Conversely, negatively connoted words are rare (Breeze 2013: 97). When faced with the need to discuss adverse circumstances or negative results, companies tend to use euphemisms such as the word challenging (Breeze 2013: 97), which frames a problem as transitory, and as one that can be solved. Interestingly, annual reports published by poorly performing companies have been found to show a stronger positive bias (Rutherford 2005), which lends support to the hypothesis that annual reports are used as vehicles for impression management to control and manipulate the readers’ perceptions of the company (Merkl-Davies and Brennan 2007). The language of annual reports is, however, not overtly positive in all sections. In this respect, annual reports can be generally subdivided into two clearly defined halves (Breeze 2013: 88). The first part of the report, which includes the review of the year and primarily fulfills a promotional function, is characterized by a more positive tone and by a greater use of pictures (Breeze 2013: 88). The second part, which contains information about the company’s financial records, generally contains fewer pictures, frequent hedging expressions, and technical language (Breeze 2013: 100). 3 CSR REPORTS The concept of CSR, ‘whereby companies integrate social and environmental concerns in their business operations and in their interaction with their stakeholders on a voluntary basis’ (European Commission 2001), has received major attention in recent years from scholars, business practitioners, and policy makers, and has become a critical concern for companies (for a diachronic overview and definitions, see Carroll 1999 and Dahlsrud 2006). One of the direct results of the growing importance of CSR is the increasing amount of information that organizations provide about their social and environmental performance (Gray et al. 2001; Bebbington et al. 2008; Milne et al. 2009). Traditionally, social and environmental disclosures have been included in annual reports (Gray et al. 1996). In addition to that, it is nowadays common for companies to issue a dedicated CSR report (KPMG 2013). CSR reports are normally published once a year and made available on the company’s Web site. CSR reports are voluntary disclosures; differently from annual reports, they are not subject to any specific legal requirements (Ditlev-Simonsen and Wenstøp 2012). Thus, even though standardization is advancing, thanks to the growing acceptance of international reporting guidelines such as those promoted by the Global Reporting Initiative, companies have considerable discretion over their content and form. CSR reports contain information on a variety of environmental, social, and corporate governance issues, such as energy use, pollution, diversity, health and safety, and human rights (Favotto et al. 2016). The range of topics addressed has broadened over the years, reflecting an increasingly rich and multifaceted conceptualization of CSR (Catenaccio 2011). The relative emphasis companies place on different topics has also changed over time, reflecting the dynamic and evolving nature of CSR reporting (Jaworska and Nanda 2016). The modalities through which the information is presented are diverse; CSR reports typically combine narrative texts, quantitative data, and multimodal elements such as graphs and pictures. The labels used for this type of documents also vary. Commonly used descriptors are corporate social responsibility report, corporate citizenship report, and sustainability report. Whereas annual reports primarily target a specialized readership of ‘experts’, CSR reports address a more diverse range of stakeholders, including customers, employees, and regular citizens (Filimon 2009). Compared to annual reports, the intended readership is not only broader, but also possibly less inclined to read the report in a careful and systematic way (Filimon 2009: 134). According to Filimon (2009: 134), the ideal reader of CSR reports is someone ‘with less specific goals towards the discourse under scrutiny, and with a propensity to a less effortful, but heuristic elaboration of the content’. Purportedly, CSR reports aim to provide a comprehensive factual account of a company’s strategies and accomplishments toward improved social and environmental sustainability. However, like the annual report, the CSR report can be considered a hybrid discourse genre, which combines informative and promotional elements (Malavasi 2011; Bondi 2016). By reporting on their commitment and actions toward greater sustainability, companies seek to demonstrate that they care for the environment and for society, and promote a ‘green’, socially responsible image of their organization (Skulstad 2002, 2008; Buhr and Reiter 2006). In this sense, it has been observed that the language used in CSR reports is often explicitly promotional. Livesey and Kearins (2002: 246), for example, note the frequent use of emotion words such as ‘caring’, ‘wanting’, ‘striving’, ‘determination’, and ‘pride’, as well as ‘metaphors of the heart’ to portray the reporting corporation in a positive light. In a similar vein, Goletani (2011: 270) discusses the use of rhetorical features that are typical of advertising discourse, such as superlatives and positive evaluative expressions. Breeze (2013: 167) highlights the use of personal narratives and photographs as persuasive devices in these texts. From a critical perspective, CSR reports have been seen as tools for social legitimation. According to proponents of legitimacy theory (for an overview, see Deegan 2014), social and environmental disclosures are employed by companies for strategic reasons, with the ultimate goal of influencing public perceptions of their organizations and legitimate their activities and interests. The basic idea behind this approach is that, to continue to exist and operate, organizations need to show that their behavior is compatible with the interests and norms of the broader social system (Dowling and Pfeffer 1975; O’Donovan 2002). CSR reports are seen as an important vehicle through which companies pursue this goal. By providing evidence of their socially responsible behavior, companies seek to establish alignment between their actions and society’s values. The image enhancing and legitimacy building functions of CSR reports are intimately connected. Fuoli (2012), for example, shows how companies operating in different industries discursively respond to the specific legitimacy challenges they face by emphasizing different aspects of their corporate identity. Hart (2014: 47–59) discusses the legitimating function of evaluative language in the CSR reports published in 2009 by Coca Cola, Nike, and Nestlé. He concludes that evaluative resources are ‘exploited in corporate social reporting to construct an identity according to which the corporation exists in a position of alignment with the values and expectations of target audiences’ (Hart 2014: 52). Bondi (2016) examines the role of forward-looking statements as image building and legitimizing devices in CSR reports. References to the future are shown to play an important part in the discursive construction of a positive corporate identity and in the discursive legitimation of organizations, as they are repeatedly used by companies to foreground their expertise and commitment to ethical values. 4 STANCE As discussed above, previous research has shown that evaluative language plays an important role in how companies promote themselves in annual and CSR reports and seek to foster a positive relationship with their stakeholders. This study compares the use of evaluative language expressions in annual and CSR reports with the aim to explore how companies discursively shape their corporate identity to meet the expectations of the specific target audiences of these texts and gain their trust. The study adopts the framework for the analysis of evaluative language in discourse developed by Biber and colleagues, and known under the label of stance (Biber and Finegan 1988, 1989; Biber et al. 1999; Conrad and Biber 2000; Biber 2004, 2006a, 2006b). Stance is an umbrella term that encompasses a variety of linguistic devices used by speakers to express ‘personal feelings, attitudes, value judgments, or assessments’ (Biber et al. 1999: 966). Stance devices can be grouped into three main semantic/functional categories: (i) attitudinal stance, (ii) epistemic stance, and (iii) style-of-speaking stance (Biber et al. 1999; Conrad and Biber 2000). Attitudinal stance concerns expressions of personal attitudes, emotions, feelings, and value judgments. It groups together linguistic devices that have been considered under alternative labels such as affect (Ochs and Schiefelin 1989), evaluation (Hunston and Thompson 2000), and attitude (Martin and White 2005). Example (1) is a case of attitudinal stance from the corpus under study. (1) I am happy to report that we also made major strides in those aspects of our strategy in 2011 […]. (Shell 2011 annual rep.) Epistemic stance refers to a speaker’s comments on the certainty, reliability, or limitations of a proposition. This category includes linguistic devices that have been studied, within alternative theoretical frameworks, under the headings of epistemic modality (Palmer 2001), evidentiality (Chafe and Nichols 1986), and hedging (Hyland 1996). Example (2) illustrates the use of epistemic stance in the corpus. (2) We believe that our compensation policies and practices appropriately balance risks and rewards […]. (Bank of America 2011 annual rep.) Finally, style-of-speaking stance refers to a speaker’s comments on the communication itself, and on the manner in which the information is presented. Example (3) is an instance of this type of stance from the corpus. (3) In short, we are part of our communities in every way possible from the largest countries to the smallest towns. (JP Morgan 2011 annual rep.) In addition to the linguistic resources listed above, Biber and Finegan (1989) and Biber 2006a, 2006b) also include modals (e.g. may, must, will) and semi-modals (e.g. be going to, ought to) under the general heading of stance. Modals and semi-modals can be subdivided into three main categories, based on their primary meanings: (i) permission/possibility/ability, (ii) obligation/necessity, and (iii) volition/prediction (Biber et al. 1999: 485). Modals and semi-modals may carry either an intrinsic or extrinsic meaning (Biber et al. 1999: 485). The former relates to meanings of permission, ability, obligation, and volition, that is, actions or events over which humans have a direct control (Biber et al. 1999: 485). Extrinsic modality refers to assessments of likelihood (Biber et al. 1999: 485). Thus, when modals are used in the latter sense, they are functionally similar to epistemic stance markers. Examples (4a) and (4b) illustrate the use of the modal should in its intrinsic and extrinsic sense, respectively. (4) (a) At JPMorgan Chase, we believe that customers should be treated like we would want to be treated ourselves. (JP Morgan 2011 annual rep.)   (b) With the economy slowly but steadily improving, we believe this trend should continue in 2012. (Bank of America 2011 annual rep.) The next section discusses the choices made in operationalizing the framework and provides corpus examples of all the categories and stance constructions considered. 4.1 Operationalizing stance The analysis presented below is based on a slightly adapted version of the framework for the analysis of stance presented in Biber (2006b: Chapter 5). The framework comprises three main categories of stance devices: attitudinal stance, epistemic stance, and modality. Markers of style-of-speaking stance were not considered, as they proved to be extremely infrequent in the corpus. As in Biber (2006b), modality was kept separate from epistemic stance, even though modals and semi-modals may convey epistemic meanings when used in the extrinsic sense, and could therefore, in principle, be assimilated to the category of epistemic stance. Following Biber (2006a, 2006b) and Biber et al. (1999), this study focuses exclusively on instances where a stance expression takes scope over a proposition, which the authors label grammatical stance. Affective or evaluative words when not expressing a stance relative to another proposition—lexical markers of stance in Biber and colleagues’ terminology—were excluded from the analysis (see Biber et al. 1999: 968). The following examples show a grammatical marker of stance and a lexical marker of stance, in turn. (5) (a) We are confident [that we have the right model to improve R&D productivity and returns]. (GSK 2011 annual rep.)   (b) This year we have made good progress. (GSK 2011 CSR rep.) In (5a), the adjective confident conveys the speaker’s stance toward the proposition embedded in the complement clause over which it takes scope (marked in square brackets). Conversely, the evaluative adjective good in (5b) is used to evaluate the reporting company’s progress, rather than some other proposition. The analysis was restricted to grammatical stance for three main reasons. First, it is the most overt and explicit strategy for expressing stance, given that a distinct grammatical structure—for example, the matrix clause in (5a)—is specifically intended for this function (Biber 2006a: 99). Lexical expressions of stance depend to a greater extent on the context and shared background between speaker and hearer for their interpretation, and are thus more susceptible to different readings (Biber et al. 1999: 969). An analysis of grammatical stance devices thus provides the most direct and unambiguous evidence for the speakers’ opinions and attitudes. Second, since grammatical stance markers take scope over propositions, by analyzing this type of constructions we can isolate direct attempts by companies to frame the reading of the information provided in the reports, and thus to influence the readers’ perceptions. Third, because of their context-dependent nature, lexical markers of stance are more difficult to identify and quantify reliably. Accordingly, analyzing this type of expressions requires resource intensive annotation and reliability testing (Fuoli 2012; Fuoli and Hommerberg 2015). Given the relatively large size of the corpus used in this study, this type of analysis would have required time and resources far beyond those available. Finally, only the instances where the expressed stance could unproblematically be attributed to the reporting company or any of its members were taken into account (see Biber et al. 1999: 968). Instances where attitudes and assessments were attributed to third parties were ignored. As far as modality is concerned, only central modal auxiliary verbs were included in the analysis; semi-modals were not considered, as they proved to be very infrequent in the corpus (see Biber et al. 1999: 483ff). Similar to Biber and Finegan (1989), Biber (2006a), and Biber (2006b), the quantitative analysis of modal verbs was based on form rather than meaning, that is, no distinction was made between extrinsic and intrinsic uses of the verbs when their frequency in the corpus was calculated. This choice was primarily motived by the fact that modals are often semantically vague. It is therefore very difficult to produce objective and reliable analyses of the function of these verbs. Example (6) demonstrates the vague use of can. (6) The majority of market inputs are actively quoted and can be validated through external sources including brokers, market transactions, and third-party pricing services. (Bank of America 2011 annual rep.) In (6), it is unclear whether can is used to express possibility, in the sense that ‘it is possible to validate market inputs through external sources’, or permission, meaning that ‘it is legally permitted’ to do so. Given the high number of ambiguous instances found in a preliminary analysis of the data, I decided not to pursue a fine-grained functional analysis of modals and adopt the polyfunctional, form-based categories used by Biber (2006b) instead (see Table 1), so as to avoid the risk of reporting unreliable results. Table 1: List of stance constructions included in the analysis Stance type Construction type Illustrative corpus examplea Source Attitudinal stance Ability or willingness adjective + complement clauseb We are determined [to continue providing leadership within our sphere of influence]. Neslté 2011 CSR rep. Attitude adverb Amazingly, [covered bonds in Europe count as 100% liquid assets…]. JP Morgan 2011 annual rep. Attitude verb + complement clauseb,c I regret [that six people died working for Shell in 2011]. Shell 2011 annual rep. Desire/intention/decision verb + to-complement clause Merck wants [to offer its talented employees the opportunity to have an interesting career]. Merck 2011 CSR rep. Ease or difficulty adjective + to-complement claused It is inherently difficult [to predict the outcome of the regulatory proceedings involving our businesses]. HSBC 2011 annual rep. Emotion/attitude adjective + complement clauseb,c I am happy [to report that we also made major strides in those aspects of our strategy in 2011…]. Shell 2011 annual rep. Evaluation adjective + complement clauseb,c,d It is essential [that we meet consistently high ethical and quality standards for research and development in all parts of our business…]. GSK 2011 CSR rep. Epistemic stance Certainty/likelihood adjective + complement clauseb,c,d With stronger capital, liquidity and reserves, we are confident [that we will continue to succeed in the ever-changing global business environment]. Bank of America 2011 CSR rep. Certainty/likelihood adverb Real estate was not the only culprit in the recent crisis, but [it certainly was at the eye of the storm]. JP Morgan 2011 annual rep. Certainty/likelihood noun + that-complement clause It remains our strong belief [that operating in a responsible and ethical way is essential for the success of GSK]. GSK 2011 annual rep. Certainty/likelihood verb + complement clauseb,c Our analysis suggests [biofuels could make up as much as 23% of global incremental demand for transport fuels over the period 2010–2030]. BP 2011 CSR rep. Modality Permission/possibility/ability modal A bigger challenge may [be convincing women that they need to be concerned about heart health issues]. Abbott 2011 CSR rep. Necessity/obligation modal These suppliers must [comply with the expectations and standards of our code…]. Kellogg 2011 CSR rep. Volition/prediction modal We will [source all sunflower oil sustainably by 2020]. Unilever 2011 CSR rep. Stance type Construction type Illustrative corpus examplea Source Attitudinal stance Ability or willingness adjective + complement clauseb We are determined [to continue providing leadership within our sphere of influence]. Neslté 2011 CSR rep. Attitude adverb Amazingly, [covered bonds in Europe count as 100% liquid assets…]. JP Morgan 2011 annual rep. Attitude verb + complement clauseb,c I regret [that six people died working for Shell in 2011]. Shell 2011 annual rep. Desire/intention/decision verb + to-complement clause Merck wants [to offer its talented employees the opportunity to have an interesting career]. Merck 2011 CSR rep. Ease or difficulty adjective + to-complement claused It is inherently difficult [to predict the outcome of the regulatory proceedings involving our businesses]. HSBC 2011 annual rep. Emotion/attitude adjective + complement clauseb,c I am happy [to report that we also made major strides in those aspects of our strategy in 2011…]. Shell 2011 annual rep. Evaluation adjective + complement clauseb,c,d It is essential [that we meet consistently high ethical and quality standards for research and development in all parts of our business…]. GSK 2011 CSR rep. Epistemic stance Certainty/likelihood adjective + complement clauseb,c,d With stronger capital, liquidity and reserves, we are confident [that we will continue to succeed in the ever-changing global business environment]. Bank of America 2011 CSR rep. Certainty/likelihood adverb Real estate was not the only culprit in the recent crisis, but [it certainly was at the eye of the storm]. JP Morgan 2011 annual rep. Certainty/likelihood noun + that-complement clause It remains our strong belief [that operating in a responsible and ethical way is essential for the success of GSK]. GSK 2011 annual rep. Certainty/likelihood verb + complement clauseb,c Our analysis suggests [biofuels could make up as much as 23% of global incremental demand for transport fuels over the period 2010–2030]. BP 2011 CSR rep. Modality Permission/possibility/ability modal A bigger challenge may [be convincing women that they need to be concerned about heart health issues]. Abbott 2011 CSR rep. Necessity/obligation modal These suppliers must [comply with the expectations and standards of our code…]. Kellogg 2011 CSR rep. Volition/prediction modal We will [source all sunflower oil sustainably by 2020]. Unilever 2011 CSR rep. Note: aThe stance marker is highlighted in bold and the proposition over which it takes scope is enclosed within square brackets. b Only complement clauses occurring in post-predicate position were taken into account. All possible structural types of complement clause were considered, that is, that-clauses, wh-clauses, infinitive clauses, and ing-clauses (see Biber et al. 1999: Chapter 9). c Instances where the complementizer that was omitted were also considered (see Biber et al. 1999: 680–3). d Extraposed constructions were also considered (see Biber et al. 1999: 660). Table 1: List of stance constructions included in the analysis Stance type Construction type Illustrative corpus examplea Source Attitudinal stance Ability or willingness adjective + complement clauseb We are determined [to continue providing leadership within our sphere of influence]. Neslté 2011 CSR rep. Attitude adverb Amazingly, [covered bonds in Europe count as 100% liquid assets…]. JP Morgan 2011 annual rep. Attitude verb + complement clauseb,c I regret [that six people died working for Shell in 2011]. Shell 2011 annual rep. Desire/intention/decision verb + to-complement clause Merck wants [to offer its talented employees the opportunity to have an interesting career]. Merck 2011 CSR rep. Ease or difficulty adjective + to-complement claused It is inherently difficult [to predict the outcome of the regulatory proceedings involving our businesses]. HSBC 2011 annual rep. Emotion/attitude adjective + complement clauseb,c I am happy [to report that we also made major strides in those aspects of our strategy in 2011…]. Shell 2011 annual rep. Evaluation adjective + complement clauseb,c,d It is essential [that we meet consistently high ethical and quality standards for research and development in all parts of our business…]. GSK 2011 CSR rep. Epistemic stance Certainty/likelihood adjective + complement clauseb,c,d With stronger capital, liquidity and reserves, we are confident [that we will continue to succeed in the ever-changing global business environment]. Bank of America 2011 CSR rep. Certainty/likelihood adverb Real estate was not the only culprit in the recent crisis, but [it certainly was at the eye of the storm]. JP Morgan 2011 annual rep. Certainty/likelihood noun + that-complement clause It remains our strong belief [that operating in a responsible and ethical way is essential for the success of GSK]. GSK 2011 annual rep. Certainty/likelihood verb + complement clauseb,c Our analysis suggests [biofuels could make up as much as 23% of global incremental demand for transport fuels over the period 2010–2030]. BP 2011 CSR rep. Modality Permission/possibility/ability modal A bigger challenge may [be convincing women that they need to be concerned about heart health issues]. Abbott 2011 CSR rep. Necessity/obligation modal These suppliers must [comply with the expectations and standards of our code…]. Kellogg 2011 CSR rep. Volition/prediction modal We will [source all sunflower oil sustainably by 2020]. Unilever 2011 CSR rep. Stance type Construction type Illustrative corpus examplea Source Attitudinal stance Ability or willingness adjective + complement clauseb We are determined [to continue providing leadership within our sphere of influence]. Neslté 2011 CSR rep. Attitude adverb Amazingly, [covered bonds in Europe count as 100% liquid assets…]. JP Morgan 2011 annual rep. Attitude verb + complement clauseb,c I regret [that six people died working for Shell in 2011]. Shell 2011 annual rep. Desire/intention/decision verb + to-complement clause Merck wants [to offer its talented employees the opportunity to have an interesting career]. Merck 2011 CSR rep. Ease or difficulty adjective + to-complement claused It is inherently difficult [to predict the outcome of the regulatory proceedings involving our businesses]. HSBC 2011 annual rep. Emotion/attitude adjective + complement clauseb,c I am happy [to report that we also made major strides in those aspects of our strategy in 2011…]. Shell 2011 annual rep. Evaluation adjective + complement clauseb,c,d It is essential [that we meet consistently high ethical and quality standards for research and development in all parts of our business…]. GSK 2011 CSR rep. Epistemic stance Certainty/likelihood adjective + complement clauseb,c,d With stronger capital, liquidity and reserves, we are confident [that we will continue to succeed in the ever-changing global business environment]. Bank of America 2011 CSR rep. Certainty/likelihood adverb Real estate was not the only culprit in the recent crisis, but [it certainly was at the eye of the storm]. JP Morgan 2011 annual rep. Certainty/likelihood noun + that-complement clause It remains our strong belief [that operating in a responsible and ethical way is essential for the success of GSK]. GSK 2011 annual rep. Certainty/likelihood verb + complement clauseb,c Our analysis suggests [biofuels could make up as much as 23% of global incremental demand for transport fuels over the period 2010–2030]. BP 2011 CSR rep. Modality Permission/possibility/ability modal A bigger challenge may [be convincing women that they need to be concerned about heart health issues]. Abbott 2011 CSR rep. Necessity/obligation modal These suppliers must [comply with the expectations and standards of our code…]. Kellogg 2011 CSR rep. Volition/prediction modal We will [source all sunflower oil sustainably by 2020]. Unilever 2011 CSR rep. Note: aThe stance marker is highlighted in bold and the proposition over which it takes scope is enclosed within square brackets. b Only complement clauses occurring in post-predicate position were taken into account. All possible structural types of complement clause were considered, that is, that-clauses, wh-clauses, infinitive clauses, and ing-clauses (see Biber et al. 1999: Chapter 9). c Instances where the complementizer that was omitted were also considered (see Biber et al. 1999: 680–3). d Extraposed constructions were also considered (see Biber et al. 1999: 660). Table 1 provides a complete list of the stance constructions surveyed. Each construction is accompanied by an illustrative example taken from the corpus. 5 THE CORPUS For this study, a specialized corpus of roughly 2.5 million words was compiled. It contains the 2011 annual reports and CSR reports published by a sample of publicly traded multinational corporations belonging to four different industry sectors. The sectors considered are financial services, oil and gas, pharmaceuticals, and food processing. The corpus includes the 2011 annual and CSR reports published by the fourth largest Europe- or US-based companies2 in each of these sectors (i) that released both an annual report and a stand-alone CSR report for the year considered (2011), and (ii) whose annual report contains both legally required financial data (e.g. audited financial statements) and voluntary information (e.g. letter to shareholders, operating, and financial review). The choice of sectors was guided by the aim to optimize corpus diversity and representativeness within the resources available for this study; the sectors examined include companies with different stakeholder priorities and relations, and with different strategic, economic, and CSR challenges. Table 2 provides summary information about the corpus. As shown in the table, the corpus is subdivided into two subcorpora for comparative purposes: the annual report subcorpus (henceforth, ARC) and the CSR report subcorpus (henceforth, CSR). A detailed overview of the corpus, including the word count for each text, can be found in the Supplementary Materials. Table 2: Overview of the corpus Total number of reports 32 (16 annual reports; 16 CSR reports) Total corpus size 2,433,956 Size of subcorpora ACR 1,828,029 CSR 605,927 Mean report size ACR 114,251.8 CSR 37,870.4 Sectors and companies Financial services JP Morgan Chase HSBC Wells Fargo Bank of America Oil and gas Exxon Mobil Royal Dutch Shell Chevron BP Pharmaceuticals Johnson & Johnson Merck GlaxoSmithKline Abbott Laboratories Food processing Neslté Unilever General Mills Kellogg Total number of reports 32 (16 annual reports; 16 CSR reports) Total corpus size 2,433,956 Size of subcorpora ACR 1,828,029 CSR 605,927 Mean report size ACR 114,251.8 CSR 37,870.4 Sectors and companies Financial services JP Morgan Chase HSBC Wells Fargo Bank of America Oil and gas Exxon Mobil Royal Dutch Shell Chevron BP Pharmaceuticals Johnson & Johnson Merck GlaxoSmithKline Abbott Laboratories Food processing Neslté Unilever General Mills Kellogg Table 2: Overview of the corpus Total number of reports 32 (16 annual reports; 16 CSR reports) Total corpus size 2,433,956 Size of subcorpora ACR 1,828,029 CSR 605,927 Mean report size ACR 114,251.8 CSR 37,870.4 Sectors and companies Financial services JP Morgan Chase HSBC Wells Fargo Bank of America Oil and gas Exxon Mobil Royal Dutch Shell Chevron BP Pharmaceuticals Johnson & Johnson Merck GlaxoSmithKline Abbott Laboratories Food processing Neslté Unilever General Mills Kellogg Total number of reports 32 (16 annual reports; 16 CSR reports) Total corpus size 2,433,956 Size of subcorpora ACR 1,828,029 CSR 605,927 Mean report size ACR 114,251.8 CSR 37,870.4 Sectors and companies Financial services JP Morgan Chase HSBC Wells Fargo Bank of America Oil and gas Exxon Mobil Royal Dutch Shell Chevron BP Pharmaceuticals Johnson & Johnson Merck GlaxoSmithKline Abbott Laboratories Food processing Neslté Unilever General Mills Kellogg 5.1 Corpus compilation and text conversion The corpus compilation process involved three main steps. First, the pdf reports were downloaded from the respective companies’ Web sites. A complete list of the URLs from which the reports were downloaded is provided in the Supplementary Materials. Second, the pdf reports were automatically converted to raw text with the aid of the open-source computer program PDFMiner.3 Third, each corpus file was manually inspected to identify and correct conversion errors, such as misspelled words due to incorrect decoding of text characters, interrupted sentences due to column or page breaks, misspelled hyphenated words due to incorrect decoding of line breaks. The inspection and correction of the files were carried out in a step-wise fashion with the aid of the advanced text editor TextWrangler,4 using regular expressions to systematically identify and amend the errors. A detailed description of the text inspection and cleaning procedure is provided in the Supplementary Materials, including the regular expressions used at each step of the process. 6 METHOD The procedure for the identification and quantification of stance constructions in the corpus relied on a combination of automatic techniques and manual analysis. It comprised three steps: 1. Automatic stance marker retrieval. For each of the categories included in the analysis framework, an extended list of stance markers was created based on the lists provided in Biber and Finegan (1989) and Biber (2006b). These lists were integrated with additional stance markers derived from the author’s prior work, from an informal exploratory analysis of the corpus, and by consulting a thesaurus of English. By means of a string search Perl script,5 the stance markers included in the lists were simultaneously and automatically searched for in the corpus. For each marker, all inflected forms, for instance believe, believes, believed, believing, were included in the query. The script output consisted of a tabled list of the markers from the original set that were actually found in the corpus and the frequency of each. In this way, the expressions that were not used in the texts under study could immediately be identified and discarded, simplifying the following analysis steps. The complete lists of stance markers used at this stage are provided in Table A1 (available in the Supplementary Materials).6 The frequency lists produced by the string search script are given in the Supplementary Materials. 2. Manual word-sense disambiguation by means of concordance analysis. Some of the stance expressions detected in step 1 are polysemous and may carry both stance and non-stance senses. Furthermore, they may be used as lexical, rather than grammatical, stance markers, and may be attributed to third parties. Compare, for example: (7) (a) ExxonMobil is proud to be a charter member of the Wildlife Habitat Council (WHC) since its establishment in 1988. (Exxon CSR rep.)   (b) Whatever the outcome, Adeline’s sights are set firmly on qualifying for the 2016 Olympics in Rio de Janeiro, Brazil. She explains: ‘I want to do well for my SOS family, my SOS mother and my country. They helped me to get where I am now and I want them to be proud of me.’ (HSBC CSR rep.) In example (7a) proud is used as a grammatical marker of attitudinal stance in an emotion/attitude adjective + to-complement clause construction. Conversely, in example (7b), the same word is used as a lexical marker of stance to describe a desired state of affairs, rather than to express the speaker’s attitude toward a proposition. In addition, this sentence is attributed to a person that does not belong to the company, namely, one of the beneficiaries of a community development plan launched by HSBC bank. For these reasons, (7b) should be excluded from the analysis (cf. analysis criteria in Section 4.1). Since the string search algorithm used in step one can only retrieve single words, it does not consider the linguistic context in which the stance expressions are found, nor does it incorporate any automatic word-sense disambiguation component, it may produce a high number of false positives and return unreliable results. Therefore, the second step of the analysis consisted in manually inspecting the results of the automatic search to identify and remove all irrelevant items. This was accomplished by means of concordance analysis with the aid of the corpus program AntConc (Anthony 2012). A Key Word in Context (KWIC) concordance was produced for each of the words retrieved by the search string algorithm in step 1, and exported to a spreadsheet.7 Each concordance line was manually inspected and annotated as either relevant or irrelevant based on the criteria outlined in Section 4.1. In a limited number of cases, ad hoc rules had to be formulated to deal with problematic expressions and complex co-textual dynamics. A full account of these ad hoc rules is included in the Supplementary Materials. 3. Quantification and statistical analysis. Once concordances for all the stance markers retrieved in step 1 were produced and all irrelevant entries were removed from them, the frequency of occurrence of each marker in the two subcorpora was calculated. Frequencies were normalized to allow for a comparison between the subcorpora, which are of different size. The aggregated frequency of stance markers was calculated and compared across the categories included in the analysis framework. Where relevant, pairwise log-likelihood tests (Rayson and Garside 2000) were performed to determine whether the differences observed in the frequency of stance markers between the ARC and CSR subcorpora were statistically significant, that is, not attributable to chance variation.8p-values equal to or lower than .05 were considered statistically significant. The next section summarizes the results of the analysis and discusses a number of examples from the corpus. 7 RESULTS This section presents the results of the analysis. It begins by examining the overall distribution of stance constructions in the corpus, and then proceeds to discussing the main findings for each of the three general stance categories considered, that is, attitudinal stance, epistemic stance, and modality, in more detail. Due to space constraints, the presentation of the results will focus on the most important findings. The complete quantitative results are provided in the Supplementary Materials, including the raw and normalized frequencies for each stance marker and category surveyed. 7.1 General distribution of stance constructions Figure 1 shows the overall frequency of stance constructions in the ARC and CSR subcorpora, grouped by type. The plot shows that modals are by far the most common type of stance marker. Modals are significantly more frequent in the CSR compared to the ARC (LL = 11.96, df = 1, p < .001). Attitudinal stance constructions are also significantly more frequent in the CSR than in the ARC (LL = 945.31, df = 1, p<.001). A minor, non-significant difference was found in the frequency of epistemic stance markers between the two subcorpora (LL = 0.45, df = 1, p = .50). Overall, the CSR contains a significantly higher number of stance markers in comparison to the ARC—8674.22 versus 6801.22 instances per million words (LL = 209.58, df = 1, p < .001). Figure 1: View largeDownload slide Distribution of stance types across corpora Figure 1: View largeDownload slide Distribution of stance types across corpora From these findings, we may infer that the language in CSR reports is more explicitly subjective and evaluative than in annual reports. In the former, companies are more prone to openly express their emotions, attitudes, and assessments, as evidenced by the comparatively higher frequency of stance markers in general, and by the overwhelmingly higher frequency of attitudinal expressions in particular. To gain a deeper understanding of these differences and of how stance markers are used in annual and CSR reports, however, we need to take a closer look at each stance category. The next section provides a more detailed account of the distribution and functions of attitudinal stance constructions in the corpus. 7.2 Attitudinal stance Figure 2 reports the frequency and distribution of attitudinal stance constructions in the corpus. As the plot shows, desire/intention/decision verb + to-clause and ability or willingness adjective + complement clause are the two most frequent attitudinal stance constructions among the seven surveyed. Both are more than three times as frequent in the CSR compared to the ARC. Figure 2: View largeDownload slide Distribution of attitudinal stance constructions Figure 2: View largeDownload slide Distribution of attitudinal stance constructions The most common desire/intention/decision verbs controlling a complement clause are reported in Figure 3. The graph shows that aim is the most frequent verb in this group, and is overwhelmingly more frequent in the CSR compared to the ARC. This result suggests that, in their CSR reports, companies place considerable emphasis on objectives and ambitions for the future. These relate primarily to positive social and environmental goals (87.93% of all instances of aim + to-clause), as illustrated in example (8). (8) We aim to be a good neighbour to the communities close to our projects and facilities. (Shell CSR rep.) Statements of good intentions like the one in (8) often appear to be used rhetorically in CSR reports, to demonstrate the reporting company’s commitment to the principles of CSR, and project integrity. As shown in Figure 3, other common desire/intention/decision verbs in the CSR are seek, strive, and want. Similar to aim, these verbs are often used in CSR reports to emphasize a company’s commitment to ethical and responsible conduct (88.43% of all instances), and thus to communicate integrity, as illustrated in (9). (9) We strive to maintain the highest standards of integrity and do the right thing for our customers, colleagues, shareholders, and all of our other partners. (JP Morgan CSR rep.) Figure 3: View largeDownload slide Distribution of desire/intention/decision verbs Figure 3: View largeDownload slide Distribution of desire/intention/decision verbs While all three verbs are markedly less frequent in the ARC compared to the CSR, strive stands out as particularly rare in the former subcorpus. One possible explanation for this considerable difference resides in the semantics of this verb. Strive expresses a genuine and earnest effort toward a goal. It evokes positive evaluative meanings of determination, tenacity, and humbleness. However, striving for a goal does not necessarily imply that it will be achieved. Such a non-binding declaration of intents, even if loaded with positive valence, may not be sufficient to reassure shareholders and investors that the company will live up to their expectations and turn their investments into profit. The only desire/intention/decision verb that is comparatively more frequent in annual reports is intend. In 59.41% of the occurrences, this verb is used to discuss a company’s strategic and financial plans, as shown in (10). (10) Within the Refining and Marketing segment, BP intends to divest the Texas City refinery and related assets […]. (BP annual rep.) Compared to most other verbs in this group, intend is a relatively neutral stance verb. It communicates rationality, and indexes emotionally controlled decision-making. The relatively frequent use of this verb to communicate intentions in annual reports may be interpreted as reflecting the companies’ need to show shareholders and investors that they are in control of the situation, and that their decisions are rational, rather than emotional. Further, compared to other verbs in this category, intend expresses a more definite commitment to a goal, and may thus contribute to strengthen the impression that the company can be trusted to deliver on its targets. Ability or willingness adjective + complement clause is the second most frequent attitudinal stance construction in both the subcorpora and, as noted above, it is more frequent in the CSR than in the ARC. Overall, the most frequent ability/willingness adjectives are committed, able, and dedicated. In the vast majority of cases (91.67%), the words committed and dedicated are employed in CSR reports to foreground a company’s determination in pursuing social and environmental objectives, as in (11). (11) We are dedicated to playing our part in addressing this important public health challenge, and micronutrient fortification will remain a focus in 2012. (Nestlé CSR rep.) The expression committed to is overwhelmingly more frequent in the CSR than in the ARC—308.61 compared to 29.54 occurrences per million words (LL = 294.51, df = 1, p < .001). This difference accounts for a substantial part of the difference in the frequency of ability/willingness adjective + complement clause constructions across the subcorpora. Example (12) illustrates this construction. (12) At Abbott, we are committed to meeting high ethical standards and to complying with all applicable local, national, and international laws wherever we do business. (Abbott CSR rep.) From the point of view of corporate identity, the large difference observed in the frequency of committed to constructions between the subcorpora is highly significant. Semantically, the adjective committed expresses a strong degree of willingness, and carries meanings of obligation, loyalty, and accountability. Thus, the frequent use of this word in CSR reports can be seen to reinforce the idea that the reporting companies are strongly and genuinely determined to become better corporate citizens. At the same time, however, committed is similar to strive in that it signals a firm intention to reach a goal, but leaves the possibility open that the goal might not actually be achieved. This might explain why it is not as frequently used in annual reports as in CSR reports. Among ability or willingness adjective + complement clause constructions, able + to-clause is also relatively frequent. This construction is often used in both report types to underline a company’s achievements and capabilities (56.76% of all instances), as example (13) shows. (13) Through year-end 2011, we have been able to achieve significant emission reductions […]. (JP Morgan CSR rep.) Once again, this construction is significantly more frequent in CSR reports compared to annual reports, with 95.72 and 49.23 occurrences per million words, respectively (LL = 14.63, df = 1, p < .001). 7.3 Epistemic stance Figure 4 plots the distribution of epistemic stance constructions in the corpus. As the graph shows, certainty/likelihood verbs controlling a complement clause are by far the most frequent resources used to express epistemic stance, followed by adverbs, adjectives controlling a complement clause, and nouns controlling a complement clause. Figure 4: View largeDownload slide Distribution of epistemic stance constructions Figure 4: View largeDownload slide Distribution of epistemic stance constructions Following Biber and Finegan (1989) and Biber (2006b), stance constructions can be broadly classified into those that express certainty about a proposition, and those that express doubt or likelihood. Figure 5 represents the distribution of epistemic stance constructions into these two subcategories in the ARC and CSR. Constructions expressing certainty are significantly more frequent in the CSR (LL = 100.03, df = 1, p < .001), whereas markers of likelihood are significantly more common in the ARC (LL = 48.53, df = 1, p < .001). This finding indicates that companies are generally more assertive in CSR reports, whereas they tend to adopt a more cautious, tentative stance in annual reports. Figure 5: View largeDownload slide Distribution of certainty and likelihood constructions Figure 5: View largeDownload slide Distribution of certainty and likelihood constructions A closer look at the frequency of the epistemic verbs retrieved in the subcorpora offers further insights into the use of certainty and likelihood stance expressions in the corpus, and the communicative functions they perform. As Figure 6 shows, only four certainty verbs are more frequent in the ARC than in the CSR: anticipate, conclude, confirm, and determine. These verbs point to two major communicative functions performed by annual reports, that is, (i) evaluating the performance and the financial and legal standing of a company, and (ii) anticipating future scenarios, decisions, and actions. Examples (14) and (15) illustrate these two functions in turn. (14) The review concluded the board had operated well in 2011. (BP annual rep.) (15) The Company anticipates that operating cash flows, existing credit facilities and access to the commercial paper markets will provide sufficient resources to fund operating needs in 2012. (Johnson & Johnson annual rep.) Figure 6: View largeDownload slide Distribution of certainty verbs controlling a complement clause Figure 6: View largeDownload slide Distribution of certainty verbs controlling a complement clause The relatively frequent use of these verbs in the ARC has important implications for the discursive construction of corporate identity. By providing informed assessments of the present situation and by carefully considering and anticipating future scenarios, companies not only address shareholders’ information needs, but also project readiness and competence; they show that they are in control of the situation and prepared for any contingency. As far as the CSR is concerned, the most common certainty verbs are show, know, find, and understand. The verbs show and find serve three main communicative functions in CSR reports. First, they are used to provide empirical evidence for the companies’ accomplishments related to different aspects of CSR (29.35% of all instances). In (16), for example, Bank of America showcases its progress toward reduced water consumption. (16) In 2011, we completed a two-year comprehensive inventory of our water consumption (our first ever). The review found that we consumed 3.9 billion gallons of water in 2011, down from 4.2 billion gallons in 2010. (Bank of America CSR rep.) Second, these verbs are used to present scientifically credible evidence for the companies’ positive contribution to the solution of pressing social and environmental problems (26.09% of all instances). In (17), for instance, the cereal company General Mills refers to research demonstrating the potential positive effects of cereals against obesity, thus suggesting that the company’s business is inherently good. (17) A 2009 study showed that regardless of sweetness level, cereal eaters have healthier body weights than those who don’t eat cereal. (General Mills CSR rep.) Third, show and find are used in CSR reports to refer to scientific research that reportedly informs the companies’ decisions and strategies (20.65% of all instances). The verbs know and understand serve an important dialogic and solidarity-building function in CSR reports; they are often used to acknowledge the stakeholders’ perspective, and to show that the reporting company is sympathetic to their concerns (63.01% of all instances). This function is exemplified in (18), where GlaxoSmithKline openly recognizes the ethical implications of its activities, thereby implicitly suggesting that the company is aware of and willing to address them. (18) We know that the research and development, manufacture and sale of our products can raise ethical issues. (GSK CSR rep.) From the point of view of corporate identity, the repeated use of know and understand as perspective taking predicates, as illustrated in (18), serves to project benevolence. By explicitly acknowledging the risks and challenges inherent in their business activities and by demonstrating sympathy and a readiness to listen to the stakeholders’ views, companies present themselves as receptive and caring organizations (cf. Livesey and Kearins 2002). Figure 7 represents the distribution of likelihood verbs across the subcorpora. The plot reveals two outstanding differences between the ARC and the CSR; the verb expect is comparatively more frequent in the ARC, whereas believe is more frequent in the CSR. These differences are statistically significant (expect: LL = 112.14, df = 1, p < .001; believe: LL = 5.23, df = 1, p = .022). The frequent use of expect in the ARC mirrors the similarly frequent use of anticipate observed above. From a corporate identity perspective, the use of these anticipatory verbs can be seen to contribute to highlight the company’s readiness for the future. Compared to anticipate, however, expect conveys a more tentative stance, and serves to hedge forecasts. Therefore, from the company’s perspective, this verb can have the twofold beneficial effect of conveying a cautious and responsible attitude toward predictions, and of cushioning the negative effects of potential misjudgments. Example (19) is an example of this verb from the corpus. (19) We are well placed in emerging markets and we expect these markets to continue to drive growth. (Unilever annual rep.) Figure 7: View largeDownload slide Distribution of likelihood verbs Figure 7: View largeDownload slide Distribution of likelihood verbs Believe is the most frequent likelihood verb in CSR reports. It is often used to foreground the reporting company’s ideological commitment to the principles of CSR (35.94% of all instances), as in (20). (20) At JPMorgan Chase, we believe our place is right at the heart of efforts to solve daunting problems like unemployment, environmental sustainability, and access to quality education. (JP Morgan CSR rep.) As Fetzer (2008) observes, believe expresses a stronger degree of epistemic commitment compared to other likelihood verbs such as think, in particular when it is preceded by a first-person pronoun, as in the example above. The frequent use of this verb to communicate the companies’ ethical values and credo thus serves to emphasize their genuine and strong commitment to the cause of CSR, and to project integrity. Believe is a very important feature of the language of annual reports as well, being the second most frequent likelihood verb. It performs three primary communicative functions in these texts. First, it is used to present management’s assessment of the company’s performance and financial situation, and of the risks it faces. This function accounts for 41.26% of the occurrences, and is illustrated in example (21). (21) We believe that our portfolio of assets remains well positioned to compete and grow value in a range of external conditions and we continue to increase both investment and operating cash. (BP annual rep.) Second, believe is frequently used to explain and justify management’s operational and strategic decisions (27.79% of all instances), as shown in (22). (22) We have focus. We do not compete in numerous categories as we believe expertise in a few businesses is preferable to having simple scale as the result of participation in many. (Kellogg annual rep.) Third, in 12.21% of the occurrences, believe is used to express opinions related to legal proceedings in which the company is involved, as shown in (23). (23) Based on the history described above, Chevron believes that this lawsuit lacks legal or factual merit. (Chevron annual rep.) A substantial proportion of the instances of believe found in annual reports occurs in the introductory message to shareholders (13.26%), where the CEO and/or Chairman review the company’s achievements, and summarize goals and expectations. In this context, believe is generally used to highlight the company’s strengths, and foster optimism about the future, as example (24) shows. (24) I am proud to be on this team and believe that our best days are ahead. (JP Morgan annual rep.) In all the cases reviewed above, believe appears to be used to promote trust in the company’s competence and ability to overcome difficulties and succeed. 7.4 Modality Figure 8 displays the distribution of modals across the three main categories in the ARC and CSR. Obligation/necessity modals are significantly more frequent in the CSR (LL = 39.43, df = 1, p < .001), permission/possibility/ability modals are significantly more common in the ARC (LL = 20.02, df = 1, p < .001), and volition/prediction modals occur significantly more often in the CSR (LL = 49.38, df = 1, p < .001). Figure 8: View largeDownload slide Distribution of modal constructions Figure 8: View largeDownload slide Distribution of modal constructions Figure 9 shows the frequency of permission/possibility/ability modals in the subcorpora. Two major differences between the ARC and the CSR can be observed: can is more than twice as frequent in the CSR than the ARC; may and could are both substantially more frequent in the ARC compared to the CSR. Figure 9: View largeDownload slide Distribution of permission/possibility/ability modals Figure 9: View largeDownload slide Distribution of permission/possibility/ability modals The comparatively higher frequency of the modals may and could in the ARC is in line with the pattern observed above for likelihood verbs, namely, that companies tend to adopt a more tentative stance in annual reports. But the relatively higher frequency of these modals also reflects the fact that one of the chief communicative functions of annual reports is that of discussing and anticipating potential, future scenarios. Not surprisingly, among the most frequent verbs appearing in the immediate-right collocate slot for may in the ARC, which are shown in Table 3, we find result, require, become, affect, vary, and change. These verbs can be related to the implications and consequences of anticipated events for the company and its investors. Table 3: Verbal collocates of may ARC CSR Collocate T-score Collocate T-score be 25.19 be 7.98 have 8.88 have 3.29 include 8.54 affect 2.98 result 7.90 differ 2.64 require 6.37 result 2.41 differ 6.12 include 2.36 become 5.93 contact 2.21 affect 5.53 exist 1.99 vary 5.52 occur 1.99 change 4.92 lead 1.96 ARC CSR Collocate T-score Collocate T-score be 25.19 be 7.98 have 8.88 have 3.29 include 8.54 affect 2.98 result 7.90 differ 2.64 require 6.37 result 2.41 differ 6.12 include 2.36 become 5.93 contact 2.21 affect 5.53 exist 1.99 vary 5.52 occur 1.99 change 4.92 lead 1.96 Table 3: Verbal collocates of may ARC CSR Collocate T-score Collocate T-score be 25.19 be 7.98 have 8.88 have 3.29 include 8.54 affect 2.98 result 7.90 differ 2.64 require 6.37 result 2.41 differ 6.12 include 2.36 become 5.93 contact 2.21 affect 5.53 exist 1.99 vary 5.52 occur 1.99 change 4.92 lead 1.96 ARC CSR Collocate T-score Collocate T-score be 25.19 be 7.98 have 8.88 have 3.29 include 8.54 affect 2.98 result 7.90 differ 2.64 require 6.37 result 2.41 differ 6.12 include 2.36 become 5.93 contact 2.21 affect 5.53 exist 1.99 vary 5.52 occur 1.99 change 4.92 lead 1.96 Compared to may and could, can expresses a more assertive stance. The comparatively higher frequency of can in CSR reports appears to be coherent with the relatively higher frequency of certainty verbs discussed above. An inspection of the immediate-right verbal collocates of this modal verb, listed in Table 4, reveals that can is frequently accompanied in CSR reports by positively valenced verbs such as help, play (primarily used in the phrase play a role), and contribute. These expressions are used to underline the potential positive outcomes of a company’s decisions and actions, as in (25), or to discuss the potential benefits of responsible business practices, as in (26). Table 4: Verbal collocates of can ARC CSR Collocate T-score Collocate T-score be 21.99 be 12.86 vary 3.97 help 6.41 do 3.85 make 4.97 affect 3.79 have 4.89 have 3.41 read 4.86 help 3.41 do 4.09 make 3.19 play 3.83 provide 3.14 contribute 3.83 take 2.92 take 3.53 include 2.84 provide 3.35 ARC CSR Collocate T-score Collocate T-score be 21.99 be 12.86 vary 3.97 help 6.41 do 3.85 make 4.97 affect 3.79 have 4.89 have 3.41 read 4.86 help 3.41 do 4.09 make 3.19 play 3.83 provide 3.14 contribute 3.83 take 2.92 take 3.53 include 2.84 provide 3.35 Table 4: Verbal collocates of can ARC CSR Collocate T-score Collocate T-score be 21.99 be 12.86 vary 3.97 help 6.41 do 3.85 make 4.97 affect 3.79 have 4.89 have 3.41 read 4.86 help 3.41 do 4.09 make 3.19 play 3.83 provide 3.14 contribute 3.83 take 2.92 take 3.53 include 2.84 provide 3.35 ARC CSR Collocate T-score Collocate T-score be 21.99 be 12.86 vary 3.97 help 6.41 do 3.85 make 4.97 affect 3.79 have 4.89 have 3.41 read 4.86 help 3.41 do 4.09 make 3.19 play 3.83 provide 3.14 contribute 3.83 take 2.92 take 3.53 include 2.84 provide 3.35 (25) Achieving higher sales volumes can help to reduce the costs of goods which we can pass on as further price reductions […]. (GSK CSR rep.) (26) Sustainable sourcing can contribute to better incomes and livelihoods for farmers and their workers. (Unilever CSR rep.) As far as volition/prediction modals are concerned, will is comparatively more frequent in the CSR, whereas would is more frequent in the ARC, as shown in Figure 10. The modal shall is infrequent in both subcorpora. Both will and would can be used to make forecasts about the future. However, compared to would, will conveys a more assertive, confident stance. The difference in the distribution of will and would between the subcorpora seems to confirm the patterns observed for the modals may and can, namely that companies tend to show a more cautious attitude to forecasts in annual reports than in CSR reports. Figure 10: View largeDownload slide Distribution of volition/prediction modals Figure 10: View largeDownload slide Distribution of volition/prediction modals To conclude the survey of modal markers, I now turn to necessity modals. As previously noted, necessity modals are overall more frequent in the CSR. Figure 11 reveals that the most striking difference between the subcorpora concerns must. An analysis of the immediate-right verbal collocates of this modal verb, provided in Table 5, shows that it is frequently accompanied in CSR reports by verbs such as follow, adhere, comply, which are commonly used with reference to the rules, guidelines, and standards that a company’s employees, contractors, and suppliers must observe (91.67% of all instances), as illustrated in (27). (27) All employees are expected to uphold the highest ethical standards of business integrity. They must comply with all applicable laws and accurately record and track all business transactions. (Unilever CSR rep.) Table 5: Verbal collocates of must ARC CSR Collocate T-score Collocate T-score be 10.16 be 7.01 have 3.42 meet 2.60 meet 2.42 follow 2.44 prove 1.99 provide 2.15 develop 1.98 continue 1.93 make 1.95 work 1.81 provide 1.93 adhere 1.72 maintain 1.71 complete 1.72 manage 1.69 comply 1.71 state 1.68 create 1.69 ARC CSR Collocate T-score Collocate T-score be 10.16 be 7.01 have 3.42 meet 2.60 meet 2.42 follow 2.44 prove 1.99 provide 2.15 develop 1.98 continue 1.93 make 1.95 work 1.81 provide 1.93 adhere 1.72 maintain 1.71 complete 1.72 manage 1.69 comply 1.71 state 1.68 create 1.69 Table 5: Verbal collocates of must ARC CSR Collocate T-score Collocate T-score be 10.16 be 7.01 have 3.42 meet 2.60 meet 2.42 follow 2.44 prove 1.99 provide 2.15 develop 1.98 continue 1.93 make 1.95 work 1.81 provide 1.93 adhere 1.72 maintain 1.71 complete 1.72 manage 1.69 comply 1.71 state 1.68 create 1.69 ARC CSR Collocate T-score Collocate T-score be 10.16 be 7.01 have 3.42 meet 2.60 meet 2.42 follow 2.44 prove 1.99 provide 2.15 develop 1.98 continue 1.93 make 1.95 work 1.81 provide 1.93 adhere 1.72 maintain 1.71 complete 1.72 manage 1.69 comply 1.71 state 1.68 create 1.69 Figure 11: View largeDownload slide Distribution of obligation/necessity modals Figure 11: View largeDownload slide Distribution of obligation/necessity modals The frequent use of necessity modals contributes to the discursive construction of the reporting companies’ responsible corporate identity; by means of these verbs, they give prominence to their ‘moral side’ and portray themselves as authoritative, responsible supervisors (cf. Fuoli 2012). 8 DISCUSSION The results of the analysis have revealed substantial differences in the way the companies included in the sample use stance expressions to construct their corporate identity in annual and CSR reports. Table 6 summarizes the main identity-building strategies identified in the two text types, alongside the textual evidence for them. The analysis has shown that, in annual reports, the companies examined attempt to present themselves as unbiased, rational, and competent decision makers. They display an objective and emotionally detached attitude toward the information provided, as evidenced by the relatively infrequent use of stance expressions in general, and of attitudinal markers in particular. They seek to convey rationality and resolve through the relatively frequent use of the verb intend. They show a cautious and responsible attitude to forecasts and assessments, as suggested by the frequent use of likelihood verbs and the modals may and could. They project competence by providing careful assessments of organizational performance and risks and reasoned justifications for their decisions through the verb believe, and demonstrate readiness by means of verbs such as expect and anticipate, which are recurrently used to discuss future scenarios. Table 6: Summary of main findings ARC CSR Identity-building strategy Textual evidence Identity-building strategy Textual evidence Display emotionally detached, objective attitude Relatively infrequent attitudinal stance expressions Highlight the ‘human side’ of the corporation Frequent attitudinal stance expressions Display rationality and definite commitment to goals Relatively frequent use of the verb intend Display firm and genuine commitment to corporate social responsibility; project integrity Frequent expressions of intentions and desires Display cautious and responsible attitude Frequent use of likelihood epistemic stance markers Display confident attitude in the face of potential skepticism Frequent use of certainty epistemic stance markers Project readiness and competence Frequent use of anticipate and expect to discuss future scenarios; frequent use of likelihood verb believe to carefully assess organizational strengths and risks, and to justify operational and strategic decisions Display heartfelt commitment to CSR; project integrity Frequent use of likelihood verb believe to present the company’s values Display cautious and responsible attitude; project readiness Frequent use of modals may and could to discuss future scenarios Provide empirical evidence for commitment to CSR Frequent use of certainty verbs show and find Display receptive and sympathetic attitude; project benevolence Frequent use of certainty verbs know and understand as perspective-taking predicates Highlight company’s positive contribution to environmental and social progress; project integrity Frequent use of modal can accompanied be by positively connoted verbs Portray the company as attentive supervisor; project integrity Frequent use of modal must ARC CSR Identity-building strategy Textual evidence Identity-building strategy Textual evidence Display emotionally detached, objective attitude Relatively infrequent attitudinal stance expressions Highlight the ‘human side’ of the corporation Frequent attitudinal stance expressions Display rationality and definite commitment to goals Relatively frequent use of the verb intend Display firm and genuine commitment to corporate social responsibility; project integrity Frequent expressions of intentions and desires Display cautious and responsible attitude Frequent use of likelihood epistemic stance markers Display confident attitude in the face of potential skepticism Frequent use of certainty epistemic stance markers Project readiness and competence Frequent use of anticipate and expect to discuss future scenarios; frequent use of likelihood verb believe to carefully assess organizational strengths and risks, and to justify operational and strategic decisions Display heartfelt commitment to CSR; project integrity Frequent use of likelihood verb believe to present the company’s values Display cautious and responsible attitude; project readiness Frequent use of modals may and could to discuss future scenarios Provide empirical evidence for commitment to CSR Frequent use of certainty verbs show and find Display receptive and sympathetic attitude; project benevolence Frequent use of certainty verbs know and understand as perspective-taking predicates Highlight company’s positive contribution to environmental and social progress; project integrity Frequent use of modal can accompanied be by positively connoted verbs Portray the company as attentive supervisor; project integrity Frequent use of modal must Table 6: Summary of main findings ARC CSR Identity-building strategy Textual evidence Identity-building strategy Textual evidence Display emotionally detached, objective attitude Relatively infrequent attitudinal stance expressions Highlight the ‘human side’ of the corporation Frequent attitudinal stance expressions Display rationality and definite commitment to goals Relatively frequent use of the verb intend Display firm and genuine commitment to corporate social responsibility; project integrity Frequent expressions of intentions and desires Display cautious and responsible attitude Frequent use of likelihood epistemic stance markers Display confident attitude in the face of potential skepticism Frequent use of certainty epistemic stance markers Project readiness and competence Frequent use of anticipate and expect to discuss future scenarios; frequent use of likelihood verb believe to carefully assess organizational strengths and risks, and to justify operational and strategic decisions Display heartfelt commitment to CSR; project integrity Frequent use of likelihood verb believe to present the company’s values Display cautious and responsible attitude; project readiness Frequent use of modals may and could to discuss future scenarios Provide empirical evidence for commitment to CSR Frequent use of certainty verbs show and find Display receptive and sympathetic attitude; project benevolence Frequent use of certainty verbs know and understand as perspective-taking predicates Highlight company’s positive contribution to environmental and social progress; project integrity Frequent use of modal can accompanied be by positively connoted verbs Portray the company as attentive supervisor; project integrity Frequent use of modal must ARC CSR Identity-building strategy Textual evidence Identity-building strategy Textual evidence Display emotionally detached, objective attitude Relatively infrequent attitudinal stance expressions Highlight the ‘human side’ of the corporation Frequent attitudinal stance expressions Display rationality and definite commitment to goals Relatively frequent use of the verb intend Display firm and genuine commitment to corporate social responsibility; project integrity Frequent expressions of intentions and desires Display cautious and responsible attitude Frequent use of likelihood epistemic stance markers Display confident attitude in the face of potential skepticism Frequent use of certainty epistemic stance markers Project readiness and competence Frequent use of anticipate and expect to discuss future scenarios; frequent use of likelihood verb believe to carefully assess organizational strengths and risks, and to justify operational and strategic decisions Display heartfelt commitment to CSR; project integrity Frequent use of likelihood verb believe to present the company’s values Display cautious and responsible attitude; project readiness Frequent use of modals may and could to discuss future scenarios Provide empirical evidence for commitment to CSR Frequent use of certainty verbs show and find Display receptive and sympathetic attitude; project benevolence Frequent use of certainty verbs know and understand as perspective-taking predicates Highlight company’s positive contribution to environmental and social progress; project integrity Frequent use of modal can accompanied be by positively connoted verbs Portray the company as attentive supervisor; project integrity Frequent use of modal must By contrast, in CSR reports, the companies included in this study use stance resources to portray themselves as committed, honest, and caring corporate citizens. They underscore their genuine commitment to the principles of CSR through the frequent use of expressions of intentions and desires. They strive to project integrity by emphasizing their ethical values through the verb believe, by highlighting their positive contribution to social and environmental progress, and by portraying themselves as attentive supervisors through the frequent use of necessity modals in relation to the rules they expect coworkers to observe. They seek to communicate benevolence by displaying a receptive and sympathetic attitude through the frequent use of know and understand as perspective-taking predicates. They foreground their ‘human side’ by frequently expressing desires, emotions, and evaluations. Overall, then, the results of this study suggest that companies tend to profile different identities in annual and CSR reports. If we accept the premise that corporate identity can function as a rhetorical, image-building tool, we can interpret the identities expressed in these two text types as strategic self-representations aimed at maximizing the persuasive appeal of the reports vis-á-vis the specific readerships they target. As mentioned above, shareholders, the principal intended readers of annual reports, primarily seek to maximize profits and minimize financial risk (Johansen and Ellerup Nielsen 2011). Since the profitability of their investment in the company crucially depends on the management’s ability to make sound decisions, they need and expect the company to be competently and responsibly managed. The discursive identity conveyed in annual reports seems to successfully address these requirements. By representing themselves as unbiased, rational, and competent decision makers, companies effectively respond to shareholders’ expectations, thereby laying the foundations for trust. CSR reports, on the other hand, address a more diverse set of stakeholders, including both internal and external groups. External stakeholders have been shown to attach particular importance to companies’ benevolence and integrity (Pirson and Malhotra 2011). Similarly, one of the characteristics that internal stakeholders with a deep relationship with the organization, such as employees, value the most is benevolence (Pirson and Malhotra 2011). As in the case of annual reports, the corporate identity projected in CSR reports appears to effectively target the intended audiences’ specific expectations. Companies mainly emphasize their integrity and benevolence in these texts, rather than other aspects of their identity. In sum, based on the results of the analysis, we may conclude that companies discursively construct and communicate distinct identities in annual and CSR reports, which appear to be strategically designed to meet the expectations of the reports’ primary target audiences, and thus to maximize the trust-building force of the texts. The results of the analysis confirm and complement some observations made in previous research on the genres of annual and CSR reports. As far as annual reports are concerned, the findings of this study provide some evidence for the dual nature of these texts as both informative and promotional texts (Bhatia 2004; Garzone 2004). However, the persuasive work performed by annual reports appears to be rather subtle, at least in comparison to CSR reports. The lower frequency of stance markers, and in particular of attitudinal stance markers, indicates that comparatively less room is given to explicitly evaluative expressions in annual reports, and thus to direct attempts to frame the reading of the information given. CSR reports appear to be relatively more explicitly subjective and evaluative, as evidenced by the significantly higher frequency of stance markers, in particular of the attitudinal type. This difference would seem to suggest that, while both report types are ‘hybrid’ discourse genres, the promotional component is more prominent in CSR reports. The tendency observed here for companies to display a caring, benevolent identity in CSR reports had already been noted by Livesey and Kearins (2002), who analyzed the pioneering CSR reports published by Shell and The Body Shop International in 1998. This study provides additional and more extensive evidence for the importance of this discursive strategy. The results of this study also confirm the previous finding that companies place considerable emphasis on commitments and on laudable goals in CSR reports (Skulstad 2008; Catenaccio 2011; Bondi 2016). My conclusions are in line with Bondi’s (2016) interpretation of this discursive maneuver as an attempt to create a ‘consistent, caring, and collaborative image of the company’ (Bondi 2016: 75). From a genre perspective, the very high frequency of expressions of desires and intentions suggests that these are distinctive features of CSR reports, and lends support to the contention that ‘making promises’ is a central rhetorical move in the genre (Skulstad 2008; Catenaccio 2011). From a critical perspective, this finding appears to at least partially corroborate the observation made by several critical scholars that good intentions and declarations of principles are given more prominence than concrete actions and achievements in CSR reports (Laine 2005; Milne et al. 2005). Clearly, more work is needed to empirically test this hypothesis, but the patterns detected here do raise important questions about the credibility of CSR reporting and about companies’ accountability. As Bondi (2016: 76) puts it, ‘planning to “be good” can certainly be appreciated, but is also hardly disputed’. One of the main findings of the analysis is the significantly higher frequency of certainty markers in CSR reports compared to annual reports. At first sight, this pattern may appear counter-intuitive. Considering that one of the chief functions of annual reports is that of soliciting or retaining the financial support of investors—a vital resource for public companies—we may reasonably expect organizations to display more confidence in these texts compared to CSR reports, where, from a communicative point of view, the stakes are lower. However, annual reports are directed at a readership of experts, who expect companies to objectively report on their performance and achievements. Thus, the target readers may view an overly confident attitude with suspicion. CSR reports, on the other hand, address a potentially skeptical readership. Skepticism toward CSR stems from two main facts. First, corporate social performance is complex to measure, due to the lack of universally accepted definitions and assessment criteria (McWilliams et al. 2006; Turker 2009). The reliability and adequacy of the data reported in CSR reports are, therefore, potentially questionable. Second, CSR communication is considered by many as a mere self-serving PR exercise carried out by companies to ‘greenwash’ their image and reputation (Skarmeas and Leonidou 2013; Bachmann and Ingenhoff 2016). For these reasons, companies need to make an extra effort to convince the reader about the reliability of their report, and about the truthfulness of the facts and figures included in it. In this light, it is unsurprising that companies frequently use certainty verbs such as show and find to provide empirical evidence for the sincerity and strength of their commitment to CSR. The frequent use of these verbs may be interpreted as a reactive discursive strategy aimed at neutralizing public skepticism by counteracting the actual or potential negative discourses surrounding the company’s CSR activities (Fuoli and Paradis 2014). Similar to the other strategies discussed above, the use of this type of stance markers exposes the inherent dialogic nature of corporate communication, and the strategic efforts made by companies to anticipate the readers’ attitudes and reactions to manage their impressions and steer them in a favorable direction. 9 CONCLUSION This study has investigated the use of stance expressions in annual and CSR reports with the aim to shed light on the discursive strategies that companies deploy to shape their corporate identity and gain the trust of the stakeholder groups that these two genres target. The analysis has revealed significant differences in the frequency and communicative functions of stance expressions in annual and CSR reports. The findings support the view that companies profile different identities in these texts to meet the expectations of the reports’ primary target audiences and maximize the trust-building force of their communications. The results of this study contribute to shed some new light on the identity work performed by companies in their reports. They also provide novel and linguistically informed insights into the impression management tactics used by companies in annual and CSR reports, by showing the important role played by stance expressions in promoting a trustworthy corporate image. Further, the study increases our knowledge of the genres of annual and CSR reports by providing a systematic analysis of stance features, and by highlighting significant differences between these two core members of the family of business reporting genres. Finally, the study provides both linguists and business communication scholars with a robust descriptive basis for critically assessing financial and CSR reporting. In conclusion, some limitations of the study should be noted. First, the analysis is restricted to grammatical markers of stance. Future work might consider extending the analysis to include lexical markers, and possibly multimodal aspects as well. Second, due to space limitations, the distribution of stance markers within and across texts, companies and industries has not been considered. This is, however, a potentially important aspect, which would add greater depth and precision to the analysis. Third, the analysis has focused on a relatively limited set of companies and economic sectors. Future studies could take into account additional organizations and industries to gain a more comprehensive and representative picture of the stance expressions and of the image-building strategies used in annual and CSR reports. Finally, the manual procedure for disambiguating stance constructions adopted here is very time-consuming. Future studies should focus efforts on automatizing this process to expedite the analysis and improve the scalability of the method. Acknowledgements I wish to thank Carita Paradis (Lund University, Sweden) for her valuable advice and help in editing and proofreading the manuscript. I would also like to thank the Editorial Board of Applied Linguistics and three anonymous reviewers for their constructive and thought-provoking comments on earlier versions of the article. Conflict of interest statement. None declared. NOTES Footnotes 1 Throughout the article, I use the term ‘speaker’ to refer to both speakers and writers. 2 Based on the 2011 Financial Times Global 500 ranking. URL: http://im.ft-static.com/content/images/33558890-98d4-11e0-bd66-00144feab49a.pdf [Accessed 20 May 2016] 3 PDFMiner is available, free of charge, at this URL: http://www.unixuser.org/∼euske/python/pdfminer/ [Accessed 20 May 2016] 4 TextWrangler is available, free of charge, at this URL: http://www.barebones.com/products/textwrangler/ [Accessed 20 May 2016] 5 The script, which is called ‘filter_by_field.pl’, is included in the supplementary materials for the following article: M. Baroni and A. Lenci. Distributional Memory: A general framework for corpus-based semantics. 2010. Computational Linguistics 36(4): 673–721. It can be downloaded, free of charge, from this URL: http://clic.cimec.unitn.it/dm/materials/scripts.tar.gz [Accessed 20 May 2016]. Instructions on how to run the script are available on demand. 6 It should be noted that this analysis step is not strictly necessary. The analysis could start from step 2, with a manual search of all the words included in the initial list. However, given the high number of potential markers, the procedure followed in step 1 greatly simplified and expedited the analysis process. 7 The concordance lines spanned 200 words to the left and to the right of the search word. 8 The log-likelihood tests were performed using the wizard developed by Andrew Hardie (Lancaster University, UK), and available at this URL: http://corpora.lancs.ac.uk/sigtest/ [Accessed 20 May 2016]. The tests were performed on the raw frequencies. REFERENCES Anderson C. J. , Imperia G. . 1992 . ‘The corporate annual report: A photo analysis of male and female portrayals ,’ Journal of Business Communication 29 : 113 – 28 . Google Scholar Crossref Search ADS Anthony L. 2012 . Antconc (version 3.2.4) [computer software]. Waseda University. Available at http://www.antlab.sci.waseda.ac.jp/. Accessed 20 May 2016. Bachmann P. , Ingenhoff D. . 2016 . ‘Legitimacy through CSR disclosures? The advantage outweighs the disadvantages ,’ Public Relations Review 42 : 386 – 94 . Google Scholar Crossref Search ADS Beattie V. , Dhanani A. , Jones M. . 2008 . ‘Investigating presentational change in U.K. annual reports. A longitudinal perspective ,’ Journal of Business Communication 45 : 181 – 222 . Google Scholar Crossref Search ADS Bebbington J. , Larrinaga C. , Moneva J. . 2008 . ‘Corporate social reporting and reputation risk management ,’ Accounting, Auditing and Accountability Journal 21 : 337 – 61 . Google Scholar Crossref Search ADS Bednarek M. 2015 . ‘“What we contrarians already know”: Individual and communal aspects of attitudinal identity’ in Groom N. , Charles M. , Suganthi J. (eds): Corpora, Grammar and Discourse: In Honour of Susan Hunston . John Benjamins , pp. 257 – 82 . Bhatia V. 2004 . Worlds of Written Discourse: A Genre-based View . Continuum . Biber D. 2004 . ‘Historical patterns for the grammatical marking of stance: A cross-register comparison ,’ Journal of Historical Pragmatics 5 : 107 – 36 . Google Scholar Crossref Search ADS Biber D. 2006a . ‘Stance in spoken and written university registers ,’ Journal of English for Academic Purposes 5 : 97 – 116 . Google Scholar Crossref Search ADS Biber D. 2006b . University Language: A Corpus-based Study of Spoken and Written Registers . John Benjamins . Biber D. , Finegan E. . 1988 . ‘Adverbial stance types in English ,’ Discourse Processes 11 : 1 – 34 . Google Scholar Crossref Search ADS Biber D. , Finegan E. . 1989 . ‘Styles of stance in English: Lexical and grammatical marking of evidentiality and affect ,’ Text and Talk 9 : 93 – 124 . Biber D. , Johansson S. , Leech G. , Conrad S. , Finegan E. . 1999 . Longman Grammar of Spoken and Written English . Pearson Education Limited . Bondi M. 2016 . ‘The future in reports: Prediction, commitment and legitimization in CSR ,’ Pragmatics and Society 7 : 57 – 81 . Google Scholar Crossref Search ADS Breeze R. 2012 . ‘Legitimation in corporate discourse: Oil corporations after Deepwater Horizon ,’ Discourse and Society 23 : 3 – 18 . Google Scholar Crossref Search ADS Breeze R. 2013 . Corporate Discourse . Bloomsbury . Buhr N. , Reiter S. . 2006 . ‘Ideology, the environment and one world view: A discourse analysis of Noranda’s environmental and sustainable development reports’ in Freedman M. , Jaggi B. (eds): Advances in Environmental Accounting & Management , vol. 3 . Emerald Group Publishing Limited , pp. 1 – 48 . Carroll A. B. 1999 . ‘Corporate social responsibility. Evolution of a definitional construct ,’ Business and Society 38 : 268 – 95 . Google Scholar Crossref Search ADS Catenaccio P. 2011 . ‘Social and environmental reports: A diachronic perspective on an emerging genre’ in Grazone G. , Gotti M. (eds): Discourse, Communication and the Enterprise: Genres and Trends . Peter Lang , pp. 169 – 91 . Chafe W. , Nichols J. . 1986 . Evidentiality: The Linguistic Coding of Epistemology . Ablex Publishing Corporation . Conrad S. , Biber D. . 2000 . ‘Adverbial marking of stance in speech and writing’ in Hunston S. , Thompson G. (eds): Evaluation in Text: Authorial Stance and the Construction of Discourse . Oxford University Press , pp. 56 – 73 . Dahlsrud A. 2006 . ‘How corporate social responsibility is defined: An analysis of 37 definitions ,’ Corporate Social Responsibility and Environmental Management 15 : 1 – 13 . Google Scholar Crossref Search ADS David C. 2001 . ‘Mythmaking in annual reports ,’ Journal of Business and Technical Communication 15 : 195 – 222 . Google Scholar Crossref Search ADS Deegan C. 2014 . ‘An overview of legitimacy theory as applied within the social and environmental accounting literature’ in Bebbington J. , Unerman J. , O’Dweyer B. (eds): Sustainability Accounting and Accountability , 2 nd edn Routledge , pp. 248 – 72 . Ditlev-Simonsen C. D. , Wenstøp S. . 2012 . ‘Companies’ ethical commitment: An analysis of the rhetoric in CSR reports ,’ Issues in Social and Environmental Accounting 5 : 65 – 81 . Google Scholar Crossref Search ADS Ditlevsen M. G. 2012a . ‘Revealing corporate identities in annual reports ,’ Corporate Communications: An International Journal 17 : 379 – 403 . Google Scholar Crossref Search ADS Ditlevsen M. G. 2012b . ‘Telling the story of Danisco’s annual reports (1935 through 2007-2008) from a communicative perspective ,’ Journal of Business and Technical Communication 26 : 92 – 115 . Google Scholar Crossref Search ADS Dowling J. , Pfeffer J. . 1975 . ‘Organizational legitimacy: Social values and organizational behavior ,’ Pacific Sociological Review 18 : 122 – 36 . Google Scholar Crossref Search ADS Englebretson R. 2007 . Stancetaking in Discourse: Subjectivity, Evaluation, Interaction . John Benjamins . European Commission . 2001 . Green paper. Promoting a European framework for corporate social responsibility. Available at http://europa.eu/rapid/press-release_DOC-01-9_en.pdf. Accessed 20 May 2016. Favotto A. , Kollman K. , Bernhagen P. . 2016 . ‘Engaging firms: The global organisational field for corporate social responsibility and national varieties of capitalism ,’ Policy and Society 35 : 13 – 27 . Google Scholar Crossref Search ADS Fetzer A. 2008 . ‘“And I think that is a very straightforward way of dealing with it”: The communicative function of cognitive verbs in political discourse ,’ Journal of Language and Social Psychology 27 : 384 – 96 . Google Scholar Crossref Search ADS Filimon A. I. 2009 . ‘Kyosei—an example of cultural keyword argumentatively exploited in corporate reporting discourse ,’ Studies in Communication Sciences 9 : 131 – 52 . Fombrun C. 1996 . Reputation: Realizing Value from the Corporate Image . Harvard Business School Press . Freeman R. E. 1984 . Strategic Management: A Stakeholder Approach . Pitman . Friedman A. L. , Miles S. . 2006 . Stakeholders: Theory and Practice . Oxford University Press . Fuoli M. 2012 . ‘Assessing social responsibility: A quantitative analysis of Appraisal in BP’s and IKEA’s social reports ,’ Discourse and Communication 6 : 55 – 81 . Google Scholar Crossref Search ADS Fuoli M. , Hommerberg C. . 2015 . ‘Optimizing transparency, reliability and replicability: Annotation principles and inter-coder agreement in the quantification of evaluative expressions ,’ Corpora 10 : 315 – 49 . Google Scholar Crossref Search ADS Fuoli M. , Paradis C. . 2014 . ‘A model of trust-repair discourse ,’ Journal of Pragmatics 74 : 52 – 69 . Google Scholar Crossref Search ADS Garzone G. 2004 . ‘Annual company reports and CEO’s letters: Discoursal features and cultural markedness’ in Candlin C. , Gotti M. (eds): Intercultural Aspects of Specialized Communication . Peter Lang , pp. 311 – 43 . Garzone G. 2005 . ‘Letters to shareholders and chairman’s statements: Textual variability and generic integrity’ in Gillaerts P. , Gotti M. (eds): Genre Variation in Business Letters . Peter Lang , pp. 179 – 204 . Goletani L. 2011 . ‘Gazprom’s environmental report: Peculiarities of an emerging genre’, in Grazone G. , Gotti M. (eds): Discourse, Communication and the Enterprise: Genres and Trends . Peter Lang , pp. 255 – 277 . Gray R. , Javad M. , Power D. , Sinclair C. . 2001 . ‘Social and environmental disclosure and corporate characteristics: A research note and extension ,’ Journal of Business Finance and Accounting 28 : 327 – 56 . Google Scholar Crossref Search ADS Gray R. , Owen D. , Adams C. . 1996 . Accounting and Accountability: Changes and Challenges in Corporate Social and Environmental Reporting . Prentice Hall . Hallahan K. , Holtzhausen D. , Van Ruler B. , Verčič D. , Sriramesh K. . 2007 . ‘Defining strategic communication ,’ International Journal of Strategic Communication 1 : 3 – 35 . Google Scholar Crossref Search ADS Hart C. 2014 . Discourse, Grammar and Ideology: Functional and Cognitive Perspectives . Bloomsbury . Hooghiemstra R. 2000 . ‘Corporate communication and impression management–new perspectives why companies engage in corporate social reporting ,’ Journal of Business Ethics 27 : 55 – 68 . Google Scholar Crossref Search ADS Hunston S. , Thompson G. . 2000 . Evaluation in Text: Authorial Stance and the Construction of Discourse . Oxford University Press . Hyland K. 1996 . ‘Writing without conviction? Hedging in science research articles ,’ Applied Linguistics 17 : 433 – 54 . Google Scholar Crossref Search ADS Hyland K. 1998 . ‘Exploring corporate rhetoric: Metadiscourse in the CEO’s letter ,’ Journal of Business Communication 35 : 224 – 44 . Google Scholar Crossref Search ADS Jaworska S. , Nanda A. . 2016 . ‘Doing well by talking good: A topic modelling-assisted discourse study of corporate social responsibility ,’ Applied Linguistics 39 : 373 – 99 . Johansen T. S. , Ellerup Nielsen A. . 2011 . ‘Strategic stakeholder dialogues: A discursive perspective on relationship building ,’ Corporate Communications: An International Journal 16 : 204 – 17 . Google Scholar Crossref Search ADS Koller V. 2009 . ‘Corporate self-presentation and self-centredness: A case for cognitive critical discourse analysis’ in Pishwa H. (ed.): Language and Social Cognition: Expression of the Social Mind . De Gruyter , pp. 267 – 287 . KPMG . 2013 . International survey of corporate sustainability reporting. Available at http://www.kpmg.com/. Accessed 20 May 2016. Laine M. 2005 . ‘Meanings of the term ‘sustainable development’ in Finnish corporate disclosures ,’ Accounting Forum 29 : 395 – 413 . Google Scholar Crossref Search ADS Lischinsky A. 2011 . ‘In times of crisis: A corpus approach to the construction of the global financial crisis in annual reports ,’ Critical Discourse Studies 8 : 153 – 68 . Google Scholar Crossref Search ADS Livesey S. , Kearins K. . 2002 . ‘Transparent and caring corporations? A study of sustainability reports by the body shop and royal Dutch/Shell ,’ Organization and Environment 15 : 233 – 58 . Google Scholar Crossref Search ADS Malavasi D. 2010 . ‘The multifaceted nature of banks annual reports as informative, promotional and corporate communication practices’ in Allori P. Evangelisti , Garzone G. (eds): Discourse, Identities and Genres in Corporate Communication . Peter Lang , pp. 211 – 33 . Malavasi D. 2011 . ‘Doing well by doing good: A comparative analysis of Nokia’s and Ericcson’s corporate social responsibility reports’ in Grazone G. , Gotti M. (eds): Discourse, Communication and the Enterprise: Genres and Trends . Peter Lang , pp. 193 – 212 . Martin J. , White P. . 2005 . The Language of Evaluation: Appraisal in English . Palgrave Macmillan . McWilliams A. , Siegel D. S. , Wright P. M. . 2006 . ‘Corporate social responsibility: Strategic implications ,’ Journal of Management Studies 43 : 1 – 18 . Google Scholar Crossref Search ADS Merkl-Davies D. M. , Brennan N. M. . 2007 . ‘Discretionary disclosure strategies in corporate narratives: Incremental information or impression management? ,’ Journal of Accounting Literature 26 : 116 – 96 . Milne M. , Kearins K. , Walton S. . 2005 . ‘Playing with magic lanterns: The New Zealand Business Council for sustainable development and corporate triple bottom line reporting’ in Proc eedings of 4 APIRA Conference , July 4-6. Available at https://ourarchive.otago.ac.nz/handle/10523/1556. Accessed 20 May 2016. Milne M. , Tregidga H. , Walton S. . 2009 . ‘Words not actions! The ideological role of sustainable development reporting ,’ Accounting, Auditing and Accountability Journal 22 : 1211 – 57 . Google Scholar Crossref Search ADS Neu D. , Warsame H. , Pedwell K. . 1998 . ‘Managing public impressions: Environmental disclosures in annual reports ,’ Accounting, Organizations and Society 23 : 265 – 82 . Google Scholar Crossref Search ADS O’Donovan G. 2002 . ‘Environmental disclosures in the annual report: Extending the applicability and predictive power of legitimacy theory ,’ Accounting, Auditing and Accountability Journal 15 : 344 – 71 . Google Scholar Crossref Search ADS Ochs E. , Schiefelin B. . 1989 . ‘Language has a heart ,’ Text and Talk 9 : 7 – 26 . Palmer F. 2001 . Mood and Modality . Cambridge University Press . Pirson M. , Malhotra D. . 2011 . ‘Foundations of organizational trust: What matters to different stakeholders? ,’ Organization Science 22 : 1087 – 104 . Google Scholar Crossref Search ADS Rayson P. , Garside R. . 2000 . ‘Comparing corpora using frequency profiling’ in Proceedings of the Workshop on Comparing Corpora, held in Conjunction with the 38th Annual Meeting of the Association for Computational Linguistics , 1-8 October 2000, pp. 1–6. Available at http://ucrel.lancs.ac.uk/llwizard.html. Accessed 20 May 2016. Rutherford B. A. 2005 . ‘Genre analysis of corporate annual report narratives a corpus linguistics-based approach ,’ Journal of Business Communication 42 : 349 – 78 . Google Scholar Crossref Search ADS Skarmeas D. , Leonidou C. N. . 2013 . ‘When consumers doubt, watch out! The role of CSR skepticism ,’ Journal of Business Research 66 : 1831 – 8 . Google Scholar Crossref Search ADS Skulstad A. 2002 . Established and Emerging Business Genres . Norwegian Academic Press . Skulstad A. 2008 . ‘Creating a “green” image in the public sphere: Corporate environmental reports in a genre perspective’ in Wodak R. , Koller V. (eds): Handbook of Communication in the Public Sphere . De Gruyter , pp. 181 – 201 . Thompson G. , Hunston S. . 2000 . ‘Evaluation: An introduction’ in Hunston S. , Thompson G. (eds): Evaluation in Text: Authorial Stance and the Construction of Discourse . Oxford University Press , pp. 1 – 27 . Turker D. 2009 . ‘Measuring corporate social responsibility: A scale development study ,’ Journal of Business Ethics 85 : 411 – 27 . Google Scholar Crossref Search ADS Van Riel C. 1995 . Principles of Corporate Communication . Prentice Hall . Walker K. 2010 . ‘A systematic review of the corporate reputation literature: Definition, measurement, and theory ,’ Corporate Reputation Review 12 : 357 – 87 . Google Scholar Crossref Search ADS Westcott Alessandri S. 2001 . ‘Modeling corporate identity: A concept explication and theoretical explanation ,’ Corporate Communications: An International Journal 6 : 173 – 82 . Google Scholar Crossref Search ADS Appendix Table A1: List of stance markers used for automatic retrieval (step 1 of the analysis) Stance type Construction type Markers Attitudinal stance Ability or willingness adjective + post-predicate complement clause able, bound, committed, dedicated, determined, eager, unable, willing Attitude adverb alarmingly, amazingly, amusingly, annoyingly, ashamedly, astonishingly, conveniently, curiously, depressingly, disappointingly, disgustingly, disturbingly, embarrassedly, enchantingly, fortunately, frighteningly, funnily, happily, hopefully, impatiently, incredibly, inevitably, interestingly, ironically, luckily, mercifully, naturally, oddly, perplexingly, predictably, preferably, proudly, refreshingly, regretfully, regrettably, remarkably, rightly, sadly, shockingly, strangely, surprisingly, suspiciously, thankfully, tragically, unaccountably, understandably, unexpectedly, unfortunately, unhappily, unluckily, unnaturally Attitude verb + post-predicate complement clausea aggravates, agitates, alarms, amazes, amuses, annoys, astonishes, begrudge, bothers, confuses, deign, delights, despise, detest, disappoints, discourages, disgusts, dislike, dismays, distresses, disturbs, dread, embarrasses, enjoy, envy, fancy, fear, frightens, hate, hope, horrifies, interests, irritates, kills, like, loathe, love, overwhelms, pains, perplexes, perturbs, pleases, prefer, puzzles, refreshes, regret, relish, resent, rubs, saddens, scares, scorns, shocks, slays, surprises, thrills, troubles, upsets, worry Desire/intention/decision verb + to-complement clause aim, commit, crave, dedicate, intend, look, mean, necessitate, need, plan, require, seek, strive, want, wish, would like Ease or difficulty adjective + to-complement clause arduous, complex, complicated, demanding, difficult, easy, hard, impossible, impracticable, (not) possible, simple, straightforward, uncomplicated, undemanding Emotion/attitude adjective + post-predicate complement clause afraid, alarmed, amazed, ashamed, concerned, delighted, depressed, disappointed, disgusted, dismayed, dissatisfied, distressed, disturbed, eager, embarrassed, enchanted, encouraged, fascinated, frightened, furious, glad, happy, hopeful, impatient, indignant, interested, irritated, jubilant, keen, mad, merciful, overjoyed, overwhelmed, perplexed, perturbed, pleased, proud, puzzled, regretful, relieved, sad, scared, shocked, surprised, thankful, unexpected, unfortunate, unhappy, worried Evaluation adjective + post-predicate complement clause alarming, amazing, amusing, annoying, appropriate, astonishing, bad, confusing, convenient, critical, crucial, delightful, disappointing, disgusting, distressing, disturbing, embarrassing, encouraging, essential, fascinating, fortunate, frightening, fundamental, funny, good, gratifying, helpful, horrible, imperative, important, improper, inappropriate, inconceivable, incredible, inevitable, interesting, interesting, ironic, irritating, key, lucky, necessary, nice, notable, noteworthy, odd, paramount, perplexing, pleasing, positive, predictable, preferable, puzzling, reasonable, regrettable, ridiculous, scary, significant, silly, strange, surprising, suspicious, terrible, tragic, unacceptable, understandable, unexpected, unfortunate, unnatural, unsetting, vital, worrisome, worrying Epistemic stance Certainty adjective + post-predicate complement clause apparent, certain, clear, confident, convinced, definite, evident, impossible, inconceivable, incontestable, incontrovertible, indisputable, indubitable, irrefutable, manifest, not possible, obvious, patent, plain, positive, sure, true, unambiguous, unarguable, undeniable, undoubted, unmistakable, unquestionable, untrue, well-known Likelihood adjective + post-predicate complement clause alleged, arguable, conceivable, disputable, doubtful, dubious, imaginable, improbable, indefinite, likely, not certain, not clear, not sure, possible, presumable, probable, questionable, reputed, seeming, supposed, uncertain, unclear, unlikely, unsure Certainty adverb actually, admittedly, assuredly, avowedly, certainly, (in/with) certainty, clearly, (of) course, decidedly, definitely, (without) doubt, doubtlessly, evidently, (in) fact, incontrovertibly, indeed, indisputably, irrefutably, manifestly, obviously, patently, plainly, (in) reality, surely, unambiguously, unarguably, undeniably, unequivocally, unquestionably, veritably Likelihood adverb allegedly, apparently, arguably, conceivably, formally, hypothetically, ideally, likely, maybe, officially, ostensibly, outwardly, perchance, perhaps, possibly, potentially, presumably, purportedly, reportedly, reputedly, seemingly, superficially, supposedly, technically, theoretically, unlikely Certainty noun + that-complement clause certainty, conclusion, conviction, deduction, fact, finding, indication, judgment, knowledge, no doubt, observation, prediction, proof, resolution, sign Likelihood noun + that-complement clause assumption, belief, claim, doubt, estimate, feeling, guess, hypothesis, inference, probability, speculation, suspect, thought, view Certainty verb + post-predicate complement clause anticipate, ascertain, calculate, conclude, confirm, corroborate, deduce, demonstrate, determine, discern, establish, find, know, note, perceive, project, prove, ratified, realize, show, signify, understand, verify Likelihood verb + post-predicate complement clause appear, assume, believe, conjecture, consider, disbelieve, doubt, estimate, expect, feel, gather, guess, hypothesize, hypothesize, imagine, indicate, infer, postulate, presume, reckon, seem, sense, speculate, suggest, suppose, suspect, think Modality Permission/possibility/ability modal can, could, may, might Necessity/obligation modal must, should Volition/prediction modal shall, will, would Stance type Construction type Markers Attitudinal stance Ability or willingness adjective + post-predicate complement clause able, bound, committed, dedicated, determined, eager, unable, willing Attitude adverb alarmingly, amazingly, amusingly, annoyingly, ashamedly, astonishingly, conveniently, curiously, depressingly, disappointingly, disgustingly, disturbingly, embarrassedly, enchantingly, fortunately, frighteningly, funnily, happily, hopefully, impatiently, incredibly, inevitably, interestingly, ironically, luckily, mercifully, naturally, oddly, perplexingly, predictably, preferably, proudly, refreshingly, regretfully, regrettably, remarkably, rightly, sadly, shockingly, strangely, surprisingly, suspiciously, thankfully, tragically, unaccountably, understandably, unexpectedly, unfortunately, unhappily, unluckily, unnaturally Attitude verb + post-predicate complement clausea aggravates, agitates, alarms, amazes, amuses, annoys, astonishes, begrudge, bothers, confuses, deign, delights, despise, detest, disappoints, discourages, disgusts, dislike, dismays, distresses, disturbs, dread, embarrasses, enjoy, envy, fancy, fear, frightens, hate, hope, horrifies, interests, irritates, kills, like, loathe, love, overwhelms, pains, perplexes, perturbs, pleases, prefer, puzzles, refreshes, regret, relish, resent, rubs, saddens, scares, scorns, shocks, slays, surprises, thrills, troubles, upsets, worry Desire/intention/decision verb + to-complement clause aim, commit, crave, dedicate, intend, look, mean, necessitate, need, plan, require, seek, strive, want, wish, would like Ease or difficulty adjective + to-complement clause arduous, complex, complicated, demanding, difficult, easy, hard, impossible, impracticable, (not) possible, simple, straightforward, uncomplicated, undemanding Emotion/attitude adjective + post-predicate complement clause afraid, alarmed, amazed, ashamed, concerned, delighted, depressed, disappointed, disgusted, dismayed, dissatisfied, distressed, disturbed, eager, embarrassed, enchanted, encouraged, fascinated, frightened, furious, glad, happy, hopeful, impatient, indignant, interested, irritated, jubilant, keen, mad, merciful, overjoyed, overwhelmed, perplexed, perturbed, pleased, proud, puzzled, regretful, relieved, sad, scared, shocked, surprised, thankful, unexpected, unfortunate, unhappy, worried Evaluation adjective + post-predicate complement clause alarming, amazing, amusing, annoying, appropriate, astonishing, bad, confusing, convenient, critical, crucial, delightful, disappointing, disgusting, distressing, disturbing, embarrassing, encouraging, essential, fascinating, fortunate, frightening, fundamental, funny, good, gratifying, helpful, horrible, imperative, important, improper, inappropriate, inconceivable, incredible, inevitable, interesting, interesting, ironic, irritating, key, lucky, necessary, nice, notable, noteworthy, odd, paramount, perplexing, pleasing, positive, predictable, preferable, puzzling, reasonable, regrettable, ridiculous, scary, significant, silly, strange, surprising, suspicious, terrible, tragic, unacceptable, understandable, unexpected, unfortunate, unnatural, unsetting, vital, worrisome, worrying Epistemic stance Certainty adjective + post-predicate complement clause apparent, certain, clear, confident, convinced, definite, evident, impossible, inconceivable, incontestable, incontrovertible, indisputable, indubitable, irrefutable, manifest, not possible, obvious, patent, plain, positive, sure, true, unambiguous, unarguable, undeniable, undoubted, unmistakable, unquestionable, untrue, well-known Likelihood adjective + post-predicate complement clause alleged, arguable, conceivable, disputable, doubtful, dubious, imaginable, improbable, indefinite, likely, not certain, not clear, not sure, possible, presumable, probable, questionable, reputed, seeming, supposed, uncertain, unclear, unlikely, unsure Certainty adverb actually, admittedly, assuredly, avowedly, certainly, (in/with) certainty, clearly, (of) course, decidedly, definitely, (without) doubt, doubtlessly, evidently, (in) fact, incontrovertibly, indeed, indisputably, irrefutably, manifestly, obviously, patently, plainly, (in) reality, surely, unambiguously, unarguably, undeniably, unequivocally, unquestionably, veritably Likelihood adverb allegedly, apparently, arguably, conceivably, formally, hypothetically, ideally, likely, maybe, officially, ostensibly, outwardly, perchance, perhaps, possibly, potentially, presumably, purportedly, reportedly, reputedly, seemingly, superficially, supposedly, technically, theoretically, unlikely Certainty noun + that-complement clause certainty, conclusion, conviction, deduction, fact, finding, indication, judgment, knowledge, no doubt, observation, prediction, proof, resolution, sign Likelihood noun + that-complement clause assumption, belief, claim, doubt, estimate, feeling, guess, hypothesis, inference, probability, speculation, suspect, thought, view Certainty verb + post-predicate complement clause anticipate, ascertain, calculate, conclude, confirm, corroborate, deduce, demonstrate, determine, discern, establish, find, know, note, perceive, project, prove, ratified, realize, show, signify, understand, verify Likelihood verb + post-predicate complement clause appear, assume, believe, conjecture, consider, disbelieve, doubt, estimate, expect, feel, gather, guess, hypothesize, hypothesize, imagine, indicate, infer, postulate, presume, reckon, seem, sense, speculate, suggest, suppose, suspect, think Modality Permission/possibility/ability modal can, could, may, might Necessity/obligation modal must, should Volition/prediction modal shall, will, would Note: aThe verbs reported in the third person were considered as stance markers only when appearing in the construction ‘it (ADV) ____ that/to’. Conversely, for the verbs reported in the base form, all inflected forms were considered. Table A1: List of stance markers used for automatic retrieval (step 1 of the analysis) Stance type Construction type Markers Attitudinal stance Ability or willingness adjective + post-predicate complement clause able, bound, committed, dedicated, determined, eager, unable, willing Attitude adverb alarmingly, amazingly, amusingly, annoyingly, ashamedly, astonishingly, conveniently, curiously, depressingly, disappointingly, disgustingly, disturbingly, embarrassedly, enchantingly, fortunately, frighteningly, funnily, happily, hopefully, impatiently, incredibly, inevitably, interestingly, ironically, luckily, mercifully, naturally, oddly, perplexingly, predictably, preferably, proudly, refreshingly, regretfully, regrettably, remarkably, rightly, sadly, shockingly, strangely, surprisingly, suspiciously, thankfully, tragically, unaccountably, understandably, unexpectedly, unfortunately, unhappily, unluckily, unnaturally Attitude verb + post-predicate complement clausea aggravates, agitates, alarms, amazes, amuses, annoys, astonishes, begrudge, bothers, confuses, deign, delights, despise, detest, disappoints, discourages, disgusts, dislike, dismays, distresses, disturbs, dread, embarrasses, enjoy, envy, fancy, fear, frightens, hate, hope, horrifies, interests, irritates, kills, like, loathe, love, overwhelms, pains, perplexes, perturbs, pleases, prefer, puzzles, refreshes, regret, relish, resent, rubs, saddens, scares, scorns, shocks, slays, surprises, thrills, troubles, upsets, worry Desire/intention/decision verb + to-complement clause aim, commit, crave, dedicate, intend, look, mean, necessitate, need, plan, require, seek, strive, want, wish, would like Ease or difficulty adjective + to-complement clause arduous, complex, complicated, demanding, difficult, easy, hard, impossible, impracticable, (not) possible, simple, straightforward, uncomplicated, undemanding Emotion/attitude adjective + post-predicate complement clause afraid, alarmed, amazed, ashamed, concerned, delighted, depressed, disappointed, disgusted, dismayed, dissatisfied, distressed, disturbed, eager, embarrassed, enchanted, encouraged, fascinated, frightened, furious, glad, happy, hopeful, impatient, indignant, interested, irritated, jubilant, keen, mad, merciful, overjoyed, overwhelmed, perplexed, perturbed, pleased, proud, puzzled, regretful, relieved, sad, scared, shocked, surprised, thankful, unexpected, unfortunate, unhappy, worried Evaluation adjective + post-predicate complement clause alarming, amazing, amusing, annoying, appropriate, astonishing, bad, confusing, convenient, critical, crucial, delightful, disappointing, disgusting, distressing, disturbing, embarrassing, encouraging, essential, fascinating, fortunate, frightening, fundamental, funny, good, gratifying, helpful, horrible, imperative, important, improper, inappropriate, inconceivable, incredible, inevitable, interesting, interesting, ironic, irritating, key, lucky, necessary, nice, notable, noteworthy, odd, paramount, perplexing, pleasing, positive, predictable, preferable, puzzling, reasonable, regrettable, ridiculous, scary, significant, silly, strange, surprising, suspicious, terrible, tragic, unacceptable, understandable, unexpected, unfortunate, unnatural, unsetting, vital, worrisome, worrying Epistemic stance Certainty adjective + post-predicate complement clause apparent, certain, clear, confident, convinced, definite, evident, impossible, inconceivable, incontestable, incontrovertible, indisputable, indubitable, irrefutable, manifest, not possible, obvious, patent, plain, positive, sure, true, unambiguous, unarguable, undeniable, undoubted, unmistakable, unquestionable, untrue, well-known Likelihood adjective + post-predicate complement clause alleged, arguable, conceivable, disputable, doubtful, dubious, imaginable, improbable, indefinite, likely, not certain, not clear, not sure, possible, presumable, probable, questionable, reputed, seeming, supposed, uncertain, unclear, unlikely, unsure Certainty adverb actually, admittedly, assuredly, avowedly, certainly, (in/with) certainty, clearly, (of) course, decidedly, definitely, (without) doubt, doubtlessly, evidently, (in) fact, incontrovertibly, indeed, indisputably, irrefutably, manifestly, obviously, patently, plainly, (in) reality, surely, unambiguously, unarguably, undeniably, unequivocally, unquestionably, veritably Likelihood adverb allegedly, apparently, arguably, conceivably, formally, hypothetically, ideally, likely, maybe, officially, ostensibly, outwardly, perchance, perhaps, possibly, potentially, presumably, purportedly, reportedly, reputedly, seemingly, superficially, supposedly, technically, theoretically, unlikely Certainty noun + that-complement clause certainty, conclusion, conviction, deduction, fact, finding, indication, judgment, knowledge, no doubt, observation, prediction, proof, resolution, sign Likelihood noun + that-complement clause assumption, belief, claim, doubt, estimate, feeling, guess, hypothesis, inference, probability, speculation, suspect, thought, view Certainty verb + post-predicate complement clause anticipate, ascertain, calculate, conclude, confirm, corroborate, deduce, demonstrate, determine, discern, establish, find, know, note, perceive, project, prove, ratified, realize, show, signify, understand, verify Likelihood verb + post-predicate complement clause appear, assume, believe, conjecture, consider, disbelieve, doubt, estimate, expect, feel, gather, guess, hypothesize, hypothesize, imagine, indicate, infer, postulate, presume, reckon, seem, sense, speculate, suggest, suppose, suspect, think Modality Permission/possibility/ability modal can, could, may, might Necessity/obligation modal must, should Volition/prediction modal shall, will, would Stance type Construction type Markers Attitudinal stance Ability or willingness adjective + post-predicate complement clause able, bound, committed, dedicated, determined, eager, unable, willing Attitude adverb alarmingly, amazingly, amusingly, annoyingly, ashamedly, astonishingly, conveniently, curiously, depressingly, disappointingly, disgustingly, disturbingly, embarrassedly, enchantingly, fortunately, frighteningly, funnily, happily, hopefully, impatiently, incredibly, inevitably, interestingly, ironically, luckily, mercifully, naturally, oddly, perplexingly, predictably, preferably, proudly, refreshingly, regretfully, regrettably, remarkably, rightly, sadly, shockingly, strangely, surprisingly, suspiciously, thankfully, tragically, unaccountably, understandably, unexpectedly, unfortunately, unhappily, unluckily, unnaturally Attitude verb + post-predicate complement clausea aggravates, agitates, alarms, amazes, amuses, annoys, astonishes, begrudge, bothers, confuses, deign, delights, despise, detest, disappoints, discourages, disgusts, dislike, dismays, distresses, disturbs, dread, embarrasses, enjoy, envy, fancy, fear, frightens, hate, hope, horrifies, interests, irritates, kills, like, loathe, love, overwhelms, pains, perplexes, perturbs, pleases, prefer, puzzles, refreshes, regret, relish, resent, rubs, saddens, scares, scorns, shocks, slays, surprises, thrills, troubles, upsets, worry Desire/intention/decision verb + to-complement clause aim, commit, crave, dedicate, intend, look, mean, necessitate, need, plan, require, seek, strive, want, wish, would like Ease or difficulty adjective + to-complement clause arduous, complex, complicated, demanding, difficult, easy, hard, impossible, impracticable, (not) possible, simple, straightforward, uncomplicated, undemanding Emotion/attitude adjective + post-predicate complement clause afraid, alarmed, amazed, ashamed, concerned, delighted, depressed, disappointed, disgusted, dismayed, dissatisfied, distressed, disturbed, eager, embarrassed, enchanted, encouraged, fascinated, frightened, furious, glad, happy, hopeful, impatient, indignant, interested, irritated, jubilant, keen, mad, merciful, overjoyed, overwhelmed, perplexed, perturbed, pleased, proud, puzzled, regretful, relieved, sad, scared, shocked, surprised, thankful, unexpected, unfortunate, unhappy, worried Evaluation adjective + post-predicate complement clause alarming, amazing, amusing, annoying, appropriate, astonishing, bad, confusing, convenient, critical, crucial, delightful, disappointing, disgusting, distressing, disturbing, embarrassing, encouraging, essential, fascinating, fortunate, frightening, fundamental, funny, good, gratifying, helpful, horrible, imperative, important, improper, inappropriate, inconceivable, incredible, inevitable, interesting, interesting, ironic, irritating, key, lucky, necessary, nice, notable, noteworthy, odd, paramount, perplexing, pleasing, positive, predictable, preferable, puzzling, reasonable, regrettable, ridiculous, scary, significant, silly, strange, surprising, suspicious, terrible, tragic, unacceptable, understandable, unexpected, unfortunate, unnatural, unsetting, vital, worrisome, worrying Epistemic stance Certainty adjective + post-predicate complement clause apparent, certain, clear, confident, convinced, definite, evident, impossible, inconceivable, incontestable, incontrovertible, indisputable, indubitable, irrefutable, manifest, not possible, obvious, patent, plain, positive, sure, true, unambiguous, unarguable, undeniable, undoubted, unmistakable, unquestionable, untrue, well-known Likelihood adjective + post-predicate complement clause alleged, arguable, conceivable, disputable, doubtful, dubious, imaginable, improbable, indefinite, likely, not certain, not clear, not sure, possible, presumable, probable, questionable, reputed, seeming, supposed, uncertain, unclear, unlikely, unsure Certainty adverb actually, admittedly, assuredly, avowedly, certainly, (in/with) certainty, clearly, (of) course, decidedly, definitely, (without) doubt, doubtlessly, evidently, (in) fact, incontrovertibly, indeed, indisputably, irrefutably, manifestly, obviously, patently, plainly, (in) reality, surely, unambiguously, unarguably, undeniably, unequivocally, unquestionably, veritably Likelihood adverb allegedly, apparently, arguably, conceivably, formally, hypothetically, ideally, likely, maybe, officially, ostensibly, outwardly, perchance, perhaps, possibly, potentially, presumably, purportedly, reportedly, reputedly, seemingly, superficially, supposedly, technically, theoretically, unlikely Certainty noun + that-complement clause certainty, conclusion, conviction, deduction, fact, finding, indication, judgment, knowledge, no doubt, observation, prediction, proof, resolution, sign Likelihood noun + that-complement clause assumption, belief, claim, doubt, estimate, feeling, guess, hypothesis, inference, probability, speculation, suspect, thought, view Certainty verb + post-predicate complement clause anticipate, ascertain, calculate, conclude, confirm, corroborate, deduce, demonstrate, determine, discern, establish, find, know, note, perceive, project, prove, ratified, realize, show, signify, understand, verify Likelihood verb + post-predicate complement clause appear, assume, believe, conjecture, consider, disbelieve, doubt, estimate, expect, feel, gather, guess, hypothesize, hypothesize, imagine, indicate, infer, postulate, presume, reckon, seem, sense, speculate, suggest, suppose, suspect, think Modality Permission/possibility/ability modal can, could, may, might Necessity/obligation modal must, should Volition/prediction modal shall, will, would Note: aThe verbs reported in the third person were considered as stance markers only when appearing in the construction ‘it (ADV) ____ that/to’. Conversely, for the verbs reported in the base form, all inflected forms were considered. © Oxford University Press 2017 This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Applied Linguistics Oxford University Press

Building a Trustworthy Corporate Identity: A Corpus-Based Analysis of Stance in Annual and Corporate Social Responsibility Reports

Loading next page...
 
/lp/ou_press/building-a-trustworthy-corporate-identity-a-corpus-based-analysis-of-TkRdeaioj4
Publisher
Oxford University Press
Copyright
© Oxford University Press 2017
ISSN
0142-6001
eISSN
1477-450X
D.O.I.
10.1093/applin/amw058
Publisher site
See Article on Publisher Site

Abstract

Abstract This article presents a corpus-based analysis of stance in annual and corporate social responsibility (CSR) reports. It investigates how companies use stance expressions to construct and promote a positive corporate identity to gain the trust of the stakeholder groups that these texts target. The results show that companies profile distinct identities in annual and CSR reports. In annual reports, they use stance resources to portray themselves as unbiased, rational, and competent decision makers. In CSR reports, they present themselves as committed, honest, and caring corporate citizens. These discursive identities are interpreted as strategic self-representations that optimize the persuasive appeal of the reports by addressing the specific expectations of the target readerships. This study sheds some new light on the identity work performed by companies in their public discourse. It also provides novel insights into the impression management strategies used by companies in annual and CSR reports. Finally, it provides both linguists and business communication scholars with a robust descriptive basis for critically assessing financial and CSR reporting. 1 INTRODUCTION Companies use communication strategically to control and positively influence the image the public has of them, and build favorable relationships with the stakeholders on which they depend (Van Riel 1995; Hallahan et al. 2007; Merkl-Davies and Brennan 2007). One key strategy they use to pursue these communicative goals is to project a positive corporate identity (Van Riel 1995; Fombrun 1996). Corporate identity refers to ‘a firm’s strategically planned and purposeful presentation of itself’ (Westcott Alessandri 2001: 177). By communicating a positive corporate identity, business organizations seek to create a positive image in the minds of the public and, over time, develop a good reputation (Van Riel 1995; Fombrun 1996; Westcott Alessandri 2001). A good reputation, in turn, can provide numerous benefits to organizations, including increased competitiveness and profitability, and a greater ability to attract customers, investors, and talented employees (for a review, see Walker 2010). Companies have multiple stakeholders with different, and often competing, expectations and demands (Freeman 1984; Friedman and Miles 2006). Investors, for example, pursue profit maximization and risk minimization; employees primarily seek identification, safety, and job satisfaction; customers demand product quality and distinctiveness (Johansen and Ellerup Nielsen 2011). But different stakeholders also have different expectations as to how a company should be in terms of its characteristics and values; that is, they have specific expectations concerning a company’s corporate identity. For instance, investors attribute considerable importance to managerial competence, employees value transparency and benevolence, and customers prioritize technical competence and integrity (Pirson and Malhotra 2011). Companies need to understand and fulfill their key stakeholders’ expectations to gain access to vital resources and be successful (Dowling and Pfeffer 1975; Freeman 1984). This article explores how companies use discourse to manage their stakeholders’ expectations and build trust with them. More specifically, it investigates how they use stance expressions (Biber and Finegan 1989; Biber et al. 1999; Conrad and Biber 2000) to discursively construct a trustworthy corporate identity in their communications with different stakeholder groups. The term stance refers to the linguistic encoding of personal feelings, attitudes, and evaluations (Biber et al. 1999). Stance resources are directly implicated in how speakers1 communicate their identity and negotiate interpersonal relations (Thompson and Hunston 2000; Martin and White 2005; Englebretson 2007; Bednarek 2015), and are therefore assumed to play an important part in the discursive construction of corporate identity. The study focuses on two major genres of corporate public discourse, that is, the annual report and the corporate social responsibility (CSR) report. The former is primarily addressed to shareholders and investors, and includes information about a company’s financial standing. The latter is targeted at a broader and more heterogeneous audience of stakeholders, and provides information about a company’s social and environmental performance. The analysis compares the use of a broad range of stance devices in a specialized corpus of annual and CSR reports recently published by a sample of large multinational corporations. It addresses the following research questions: Q1: Are there differences in the frequency and type of stance expressions used in annual and CSR reports? Q2: How are stance expressions used to discursively construct and promote a trustworthy corporate identity in these two text types? Q3: Do companies present themselves differently in annual and CSR reports? And, if so, how do they discursively shape their identity to appeal to and earn the trust of the audiences that these texts target? The article offers several contributions to the literature on corporate identity and corporate communication. First, it extends previous discourse analysis work on corporate identity (Koller 2009; Fuoli 2012; Bondi 2016), by (i) examining the role of stance in the discursive process of corporate identity building, and (ii) investigating how companies strategically shape their discursive selves to meet the expectations of different stakeholders. Second, it provides novel, linguistically informed insights into the impression management strategies deployed by companies in their reports (Neu et al. 1998; Hooghiemstra 2000; Merkl-Davies and Brennan 2007). Third, the study extends our knowledge of the genres of annual and CSR reports by providing a systematic descriptive account of stance in these text types, and by identifying differences between them. Finally, the study contributes to the critical literature on corporate reporting (Livesey and Kearins 2002; Lischinsky 2011; Deegan 2014; Jaworska and Nanda 2016), by exposing some important persuasive—and potentially manipulative—discursive tactics used in annual and CSR reports. The article is organized as follows. Sections 2 and 3 review previous discourse analysis work on annual and CSR reports, respectively. Section 4 introduces the concept of stance and outlines the analysis framework. Section 5 describes the corpus. Section 6 discusses the method used to identify and quantify stance expressions. Section 7 presents the results of the analysis. Finally, section 8 discusses the main findings and section 9 offers conclusions. 2 ANNUAL REPORTS The annual report is a formal document that incorporates both legally required and voluntary information about a company’s financial situation and future prospects. It is explicitly addressed to the company’s shareholders, although the readership is larger, and includes public and private investors, financial experts, government authorities, specialized media, brokers, creditors, and employees (Anderson and Imperia 1992; Ditlevsen 2012b). The annual report is generally regarded as a ‘hybrid’ discourse genre that combines informative and persuasive elements (Bhatia 2004; Garzone 2004; Rutherford 2005; Malavasi 2010; Breeze 2013). It can be seen to fulfill two main communicative purposes: (i) to give a true and fair view of the state of the company’s affairs, and (ii) to provide a positive image of the company (Ditlevsen 2012b: 97). The different sections within the report play different roles in this respect. Some sections have a primarily informative function, while others are mainly devoted to promoting the company’s image (Breeze 2013: 87). In general, narrative sections containing discretionary information tend to be more overtly promotional. For example, the CEO letter, which normally introduces the report by providing an overview of the company’s performance, is typically explicitly rhetorical and optimistic (Hyland 1998; Garzone 2004, 2005; Breeze 2012; Fuoli and Paradis 2014). Conversely, sections that contain mandatory or audited financial statements have a more neutral, objective, and technical tone (Breeze 2013: 87–8). While the annual report remains ‘first and foremost a statutory document’ (Ditlevsen 2012b: 110) that is used by companies to meet their legal requirement to inform shareholders about their financial situation, diachronic research has shown that its promotional function has gained increasing prominence (Beattie et al. 2008; Ditlevsen 2012b). The annual report is now commonly used by companies as a public relations tool to build a positive corporate image and promote trust (Rutherford 2005; Beattie et al. 2008; Malavasi 2010; Ditlevsen 2012b; Breeze 2013). This transition is evidenced by the increasing number of pages, texts, and pictures included in the annual report (Beattie et al. 2008; Ditlevsen 2012b). As Beattie et al. (2008: 186) observe, the annual report has changed over time from ‘a financially driven, statutory document toward a more design-oriented document’. It has also become increasingly multimodal (David 2001; Garzone 2004; Rutherford 2005; Ditlevsen 2012a,b), and it is today a ‘complex and colorful publication’ (Breeze 2013: 85). Given the important role the report has in orienting investors’ decisions, this shift is not uncontroversial. The informative and promotional goals of the annual report can sometimes conflict with one another. Graphs and pictures, for example, have been found to be often used as impression management devices, that is, ‘to convey a more favorable impression of the organization than is warranted’ (Beattie et al. 2008: 183). Therefore, the overtly persuasive nature of the modern annual report has often been criticized (Anderson and Imperia 1992; Breeze 2013). The promotional function of annual reports is complex and multifaceted. First and foremost, through the annual report, the company wants to present itself as an attractive investment to the shareholders and potential investors, encouraging them to maintain or increase their financial support (Ditlevsen 2012b; Breeze 2013). As Ditlevsen (2012a: 379) puts it, in the annual report ‘a company displays its successes and the investment potential of its shares in order to give an impression of its ability to succeed in the future’. More broadly, the annual report is used as a vehicle for constructing and promoting a positive corporate identity (Hyland 1998; Malavasi 2010; Ditlevsen 2012a; Breeze 2013). Companies discursively ‘engineer’ a positive corporate self by emphasizing their positive attributes, aims, and values (Breeze 2013). In addition, the annual report has a relational and rapport-building function; it is used to foster good relationships with investors and other stakeholders, to encourage confidence in the company’s management team, and to build trust in the company (Malavasi 2010: 212; Breeze 2013: 84). The language used in annual reports combines purely referential elements with affective and evaluative elements. The latter tend to be more prominent in the non-technical narrative sections (Garzone 2004). Positively charged words and expressions are frequent, and are used to highlight the company’s achievements, competitiveness, and ethical values (Malavasi 2010). Conversely, negatively connoted words are rare (Breeze 2013: 97). When faced with the need to discuss adverse circumstances or negative results, companies tend to use euphemisms such as the word challenging (Breeze 2013: 97), which frames a problem as transitory, and as one that can be solved. Interestingly, annual reports published by poorly performing companies have been found to show a stronger positive bias (Rutherford 2005), which lends support to the hypothesis that annual reports are used as vehicles for impression management to control and manipulate the readers’ perceptions of the company (Merkl-Davies and Brennan 2007). The language of annual reports is, however, not overtly positive in all sections. In this respect, annual reports can be generally subdivided into two clearly defined halves (Breeze 2013: 88). The first part of the report, which includes the review of the year and primarily fulfills a promotional function, is characterized by a more positive tone and by a greater use of pictures (Breeze 2013: 88). The second part, which contains information about the company’s financial records, generally contains fewer pictures, frequent hedging expressions, and technical language (Breeze 2013: 100). 3 CSR REPORTS The concept of CSR, ‘whereby companies integrate social and environmental concerns in their business operations and in their interaction with their stakeholders on a voluntary basis’ (European Commission 2001), has received major attention in recent years from scholars, business practitioners, and policy makers, and has become a critical concern for companies (for a diachronic overview and definitions, see Carroll 1999 and Dahlsrud 2006). One of the direct results of the growing importance of CSR is the increasing amount of information that organizations provide about their social and environmental performance (Gray et al. 2001; Bebbington et al. 2008; Milne et al. 2009). Traditionally, social and environmental disclosures have been included in annual reports (Gray et al. 1996). In addition to that, it is nowadays common for companies to issue a dedicated CSR report (KPMG 2013). CSR reports are normally published once a year and made available on the company’s Web site. CSR reports are voluntary disclosures; differently from annual reports, they are not subject to any specific legal requirements (Ditlev-Simonsen and Wenstøp 2012). Thus, even though standardization is advancing, thanks to the growing acceptance of international reporting guidelines such as those promoted by the Global Reporting Initiative, companies have considerable discretion over their content and form. CSR reports contain information on a variety of environmental, social, and corporate governance issues, such as energy use, pollution, diversity, health and safety, and human rights (Favotto et al. 2016). The range of topics addressed has broadened over the years, reflecting an increasingly rich and multifaceted conceptualization of CSR (Catenaccio 2011). The relative emphasis companies place on different topics has also changed over time, reflecting the dynamic and evolving nature of CSR reporting (Jaworska and Nanda 2016). The modalities through which the information is presented are diverse; CSR reports typically combine narrative texts, quantitative data, and multimodal elements such as graphs and pictures. The labels used for this type of documents also vary. Commonly used descriptors are corporate social responsibility report, corporate citizenship report, and sustainability report. Whereas annual reports primarily target a specialized readership of ‘experts’, CSR reports address a more diverse range of stakeholders, including customers, employees, and regular citizens (Filimon 2009). Compared to annual reports, the intended readership is not only broader, but also possibly less inclined to read the report in a careful and systematic way (Filimon 2009: 134). According to Filimon (2009: 134), the ideal reader of CSR reports is someone ‘with less specific goals towards the discourse under scrutiny, and with a propensity to a less effortful, but heuristic elaboration of the content’. Purportedly, CSR reports aim to provide a comprehensive factual account of a company’s strategies and accomplishments toward improved social and environmental sustainability. However, like the annual report, the CSR report can be considered a hybrid discourse genre, which combines informative and promotional elements (Malavasi 2011; Bondi 2016). By reporting on their commitment and actions toward greater sustainability, companies seek to demonstrate that they care for the environment and for society, and promote a ‘green’, socially responsible image of their organization (Skulstad 2002, 2008; Buhr and Reiter 2006). In this sense, it has been observed that the language used in CSR reports is often explicitly promotional. Livesey and Kearins (2002: 246), for example, note the frequent use of emotion words such as ‘caring’, ‘wanting’, ‘striving’, ‘determination’, and ‘pride’, as well as ‘metaphors of the heart’ to portray the reporting corporation in a positive light. In a similar vein, Goletani (2011: 270) discusses the use of rhetorical features that are typical of advertising discourse, such as superlatives and positive evaluative expressions. Breeze (2013: 167) highlights the use of personal narratives and photographs as persuasive devices in these texts. From a critical perspective, CSR reports have been seen as tools for social legitimation. According to proponents of legitimacy theory (for an overview, see Deegan 2014), social and environmental disclosures are employed by companies for strategic reasons, with the ultimate goal of influencing public perceptions of their organizations and legitimate their activities and interests. The basic idea behind this approach is that, to continue to exist and operate, organizations need to show that their behavior is compatible with the interests and norms of the broader social system (Dowling and Pfeffer 1975; O’Donovan 2002). CSR reports are seen as an important vehicle through which companies pursue this goal. By providing evidence of their socially responsible behavior, companies seek to establish alignment between their actions and society’s values. The image enhancing and legitimacy building functions of CSR reports are intimately connected. Fuoli (2012), for example, shows how companies operating in different industries discursively respond to the specific legitimacy challenges they face by emphasizing different aspects of their corporate identity. Hart (2014: 47–59) discusses the legitimating function of evaluative language in the CSR reports published in 2009 by Coca Cola, Nike, and Nestlé. He concludes that evaluative resources are ‘exploited in corporate social reporting to construct an identity according to which the corporation exists in a position of alignment with the values and expectations of target audiences’ (Hart 2014: 52). Bondi (2016) examines the role of forward-looking statements as image building and legitimizing devices in CSR reports. References to the future are shown to play an important part in the discursive construction of a positive corporate identity and in the discursive legitimation of organizations, as they are repeatedly used by companies to foreground their expertise and commitment to ethical values. 4 STANCE As discussed above, previous research has shown that evaluative language plays an important role in how companies promote themselves in annual and CSR reports and seek to foster a positive relationship with their stakeholders. This study compares the use of evaluative language expressions in annual and CSR reports with the aim to explore how companies discursively shape their corporate identity to meet the expectations of the specific target audiences of these texts and gain their trust. The study adopts the framework for the analysis of evaluative language in discourse developed by Biber and colleagues, and known under the label of stance (Biber and Finegan 1988, 1989; Biber et al. 1999; Conrad and Biber 2000; Biber 2004, 2006a, 2006b). Stance is an umbrella term that encompasses a variety of linguistic devices used by speakers to express ‘personal feelings, attitudes, value judgments, or assessments’ (Biber et al. 1999: 966). Stance devices can be grouped into three main semantic/functional categories: (i) attitudinal stance, (ii) epistemic stance, and (iii) style-of-speaking stance (Biber et al. 1999; Conrad and Biber 2000). Attitudinal stance concerns expressions of personal attitudes, emotions, feelings, and value judgments. It groups together linguistic devices that have been considered under alternative labels such as affect (Ochs and Schiefelin 1989), evaluation (Hunston and Thompson 2000), and attitude (Martin and White 2005). Example (1) is a case of attitudinal stance from the corpus under study. (1) I am happy to report that we also made major strides in those aspects of our strategy in 2011 […]. (Shell 2011 annual rep.) Epistemic stance refers to a speaker’s comments on the certainty, reliability, or limitations of a proposition. This category includes linguistic devices that have been studied, within alternative theoretical frameworks, under the headings of epistemic modality (Palmer 2001), evidentiality (Chafe and Nichols 1986), and hedging (Hyland 1996). Example (2) illustrates the use of epistemic stance in the corpus. (2) We believe that our compensation policies and practices appropriately balance risks and rewards […]. (Bank of America 2011 annual rep.) Finally, style-of-speaking stance refers to a speaker’s comments on the communication itself, and on the manner in which the information is presented. Example (3) is an instance of this type of stance from the corpus. (3) In short, we are part of our communities in every way possible from the largest countries to the smallest towns. (JP Morgan 2011 annual rep.) In addition to the linguistic resources listed above, Biber and Finegan (1989) and Biber 2006a, 2006b) also include modals (e.g. may, must, will) and semi-modals (e.g. be going to, ought to) under the general heading of stance. Modals and semi-modals can be subdivided into three main categories, based on their primary meanings: (i) permission/possibility/ability, (ii) obligation/necessity, and (iii) volition/prediction (Biber et al. 1999: 485). Modals and semi-modals may carry either an intrinsic or extrinsic meaning (Biber et al. 1999: 485). The former relates to meanings of permission, ability, obligation, and volition, that is, actions or events over which humans have a direct control (Biber et al. 1999: 485). Extrinsic modality refers to assessments of likelihood (Biber et al. 1999: 485). Thus, when modals are used in the latter sense, they are functionally similar to epistemic stance markers. Examples (4a) and (4b) illustrate the use of the modal should in its intrinsic and extrinsic sense, respectively. (4) (a) At JPMorgan Chase, we believe that customers should be treated like we would want to be treated ourselves. (JP Morgan 2011 annual rep.)   (b) With the economy slowly but steadily improving, we believe this trend should continue in 2012. (Bank of America 2011 annual rep.) The next section discusses the choices made in operationalizing the framework and provides corpus examples of all the categories and stance constructions considered. 4.1 Operationalizing stance The analysis presented below is based on a slightly adapted version of the framework for the analysis of stance presented in Biber (2006b: Chapter 5). The framework comprises three main categories of stance devices: attitudinal stance, epistemic stance, and modality. Markers of style-of-speaking stance were not considered, as they proved to be extremely infrequent in the corpus. As in Biber (2006b), modality was kept separate from epistemic stance, even though modals and semi-modals may convey epistemic meanings when used in the extrinsic sense, and could therefore, in principle, be assimilated to the category of epistemic stance. Following Biber (2006a, 2006b) and Biber et al. (1999), this study focuses exclusively on instances where a stance expression takes scope over a proposition, which the authors label grammatical stance. Affective or evaluative words when not expressing a stance relative to another proposition—lexical markers of stance in Biber and colleagues’ terminology—were excluded from the analysis (see Biber et al. 1999: 968). The following examples show a grammatical marker of stance and a lexical marker of stance, in turn. (5) (a) We are confident [that we have the right model to improve R&D productivity and returns]. (GSK 2011 annual rep.)   (b) This year we have made good progress. (GSK 2011 CSR rep.) In (5a), the adjective confident conveys the speaker’s stance toward the proposition embedded in the complement clause over which it takes scope (marked in square brackets). Conversely, the evaluative adjective good in (5b) is used to evaluate the reporting company’s progress, rather than some other proposition. The analysis was restricted to grammatical stance for three main reasons. First, it is the most overt and explicit strategy for expressing stance, given that a distinct grammatical structure—for example, the matrix clause in (5a)—is specifically intended for this function (Biber 2006a: 99). Lexical expressions of stance depend to a greater extent on the context and shared background between speaker and hearer for their interpretation, and are thus more susceptible to different readings (Biber et al. 1999: 969). An analysis of grammatical stance devices thus provides the most direct and unambiguous evidence for the speakers’ opinions and attitudes. Second, since grammatical stance markers take scope over propositions, by analyzing this type of constructions we can isolate direct attempts by companies to frame the reading of the information provided in the reports, and thus to influence the readers’ perceptions. Third, because of their context-dependent nature, lexical markers of stance are more difficult to identify and quantify reliably. Accordingly, analyzing this type of expressions requires resource intensive annotation and reliability testing (Fuoli 2012; Fuoli and Hommerberg 2015). Given the relatively large size of the corpus used in this study, this type of analysis would have required time and resources far beyond those available. Finally, only the instances where the expressed stance could unproblematically be attributed to the reporting company or any of its members were taken into account (see Biber et al. 1999: 968). Instances where attitudes and assessments were attributed to third parties were ignored. As far as modality is concerned, only central modal auxiliary verbs were included in the analysis; semi-modals were not considered, as they proved to be very infrequent in the corpus (see Biber et al. 1999: 483ff). Similar to Biber and Finegan (1989), Biber (2006a), and Biber (2006b), the quantitative analysis of modal verbs was based on form rather than meaning, that is, no distinction was made between extrinsic and intrinsic uses of the verbs when their frequency in the corpus was calculated. This choice was primarily motived by the fact that modals are often semantically vague. It is therefore very difficult to produce objective and reliable analyses of the function of these verbs. Example (6) demonstrates the vague use of can. (6) The majority of market inputs are actively quoted and can be validated through external sources including brokers, market transactions, and third-party pricing services. (Bank of America 2011 annual rep.) In (6), it is unclear whether can is used to express possibility, in the sense that ‘it is possible to validate market inputs through external sources’, or permission, meaning that ‘it is legally permitted’ to do so. Given the high number of ambiguous instances found in a preliminary analysis of the data, I decided not to pursue a fine-grained functional analysis of modals and adopt the polyfunctional, form-based categories used by Biber (2006b) instead (see Table 1), so as to avoid the risk of reporting unreliable results. Table 1: List of stance constructions included in the analysis Stance type Construction type Illustrative corpus examplea Source Attitudinal stance Ability or willingness adjective + complement clauseb We are determined [to continue providing leadership within our sphere of influence]. Neslté 2011 CSR rep. Attitude adverb Amazingly, [covered bonds in Europe count as 100% liquid assets…]. JP Morgan 2011 annual rep. Attitude verb + complement clauseb,c I regret [that six people died working for Shell in 2011]. Shell 2011 annual rep. Desire/intention/decision verb + to-complement clause Merck wants [to offer its talented employees the opportunity to have an interesting career]. Merck 2011 CSR rep. Ease or difficulty adjective + to-complement claused It is inherently difficult [to predict the outcome of the regulatory proceedings involving our businesses]. HSBC 2011 annual rep. Emotion/attitude adjective + complement clauseb,c I am happy [to report that we also made major strides in those aspects of our strategy in 2011…]. Shell 2011 annual rep. Evaluation adjective + complement clauseb,c,d It is essential [that we meet consistently high ethical and quality standards for research and development in all parts of our business…]. GSK 2011 CSR rep. Epistemic stance Certainty/likelihood adjective + complement clauseb,c,d With stronger capital, liquidity and reserves, we are confident [that we will continue to succeed in the ever-changing global business environment]. Bank of America 2011 CSR rep. Certainty/likelihood adverb Real estate was not the only culprit in the recent crisis, but [it certainly was at the eye of the storm]. JP Morgan 2011 annual rep. Certainty/likelihood noun + that-complement clause It remains our strong belief [that operating in a responsible and ethical way is essential for the success of GSK]. GSK 2011 annual rep. Certainty/likelihood verb + complement clauseb,c Our analysis suggests [biofuels could make up as much as 23% of global incremental demand for transport fuels over the period 2010–2030]. BP 2011 CSR rep. Modality Permission/possibility/ability modal A bigger challenge may [be convincing women that they need to be concerned about heart health issues]. Abbott 2011 CSR rep. Necessity/obligation modal These suppliers must [comply with the expectations and standards of our code…]. Kellogg 2011 CSR rep. Volition/prediction modal We will [source all sunflower oil sustainably by 2020]. Unilever 2011 CSR rep. Stance type Construction type Illustrative corpus examplea Source Attitudinal stance Ability or willingness adjective + complement clauseb We are determined [to continue providing leadership within our sphere of influence]. Neslté 2011 CSR rep. Attitude adverb Amazingly, [covered bonds in Europe count as 100% liquid assets…]. JP Morgan 2011 annual rep. Attitude verb + complement clauseb,c I regret [that six people died working for Shell in 2011]. Shell 2011 annual rep. Desire/intention/decision verb + to-complement clause Merck wants [to offer its talented employees the opportunity to have an interesting career]. Merck 2011 CSR rep. Ease or difficulty adjective + to-complement claused It is inherently difficult [to predict the outcome of the regulatory proceedings involving our businesses]. HSBC 2011 annual rep. Emotion/attitude adjective + complement clauseb,c I am happy [to report that we also made major strides in those aspects of our strategy in 2011…]. Shell 2011 annual rep. Evaluation adjective + complement clauseb,c,d It is essential [that we meet consistently high ethical and quality standards for research and development in all parts of our business…]. GSK 2011 CSR rep. Epistemic stance Certainty/likelihood adjective + complement clauseb,c,d With stronger capital, liquidity and reserves, we are confident [that we will continue to succeed in the ever-changing global business environment]. Bank of America 2011 CSR rep. Certainty/likelihood adverb Real estate was not the only culprit in the recent crisis, but [it certainly was at the eye of the storm]. JP Morgan 2011 annual rep. Certainty/likelihood noun + that-complement clause It remains our strong belief [that operating in a responsible and ethical way is essential for the success of GSK]. GSK 2011 annual rep. Certainty/likelihood verb + complement clauseb,c Our analysis suggests [biofuels could make up as much as 23% of global incremental demand for transport fuels over the period 2010–2030]. BP 2011 CSR rep. Modality Permission/possibility/ability modal A bigger challenge may [be convincing women that they need to be concerned about heart health issues]. Abbott 2011 CSR rep. Necessity/obligation modal These suppliers must [comply with the expectations and standards of our code…]. Kellogg 2011 CSR rep. Volition/prediction modal We will [source all sunflower oil sustainably by 2020]. Unilever 2011 CSR rep. Note: aThe stance marker is highlighted in bold and the proposition over which it takes scope is enclosed within square brackets. b Only complement clauses occurring in post-predicate position were taken into account. All possible structural types of complement clause were considered, that is, that-clauses, wh-clauses, infinitive clauses, and ing-clauses (see Biber et al. 1999: Chapter 9). c Instances where the complementizer that was omitted were also considered (see Biber et al. 1999: 680–3). d Extraposed constructions were also considered (see Biber et al. 1999: 660). Table 1: List of stance constructions included in the analysis Stance type Construction type Illustrative corpus examplea Source Attitudinal stance Ability or willingness adjective + complement clauseb We are determined [to continue providing leadership within our sphere of influence]. Neslté 2011 CSR rep. Attitude adverb Amazingly, [covered bonds in Europe count as 100% liquid assets…]. JP Morgan 2011 annual rep. Attitude verb + complement clauseb,c I regret [that six people died working for Shell in 2011]. Shell 2011 annual rep. Desire/intention/decision verb + to-complement clause Merck wants [to offer its talented employees the opportunity to have an interesting career]. Merck 2011 CSR rep. Ease or difficulty adjective + to-complement claused It is inherently difficult [to predict the outcome of the regulatory proceedings involving our businesses]. HSBC 2011 annual rep. Emotion/attitude adjective + complement clauseb,c I am happy [to report that we also made major strides in those aspects of our strategy in 2011…]. Shell 2011 annual rep. Evaluation adjective + complement clauseb,c,d It is essential [that we meet consistently high ethical and quality standards for research and development in all parts of our business…]. GSK 2011 CSR rep. Epistemic stance Certainty/likelihood adjective + complement clauseb,c,d With stronger capital, liquidity and reserves, we are confident [that we will continue to succeed in the ever-changing global business environment]. Bank of America 2011 CSR rep. Certainty/likelihood adverb Real estate was not the only culprit in the recent crisis, but [it certainly was at the eye of the storm]. JP Morgan 2011 annual rep. Certainty/likelihood noun + that-complement clause It remains our strong belief [that operating in a responsible and ethical way is essential for the success of GSK]. GSK 2011 annual rep. Certainty/likelihood verb + complement clauseb,c Our analysis suggests [biofuels could make up as much as 23% of global incremental demand for transport fuels over the period 2010–2030]. BP 2011 CSR rep. Modality Permission/possibility/ability modal A bigger challenge may [be convincing women that they need to be concerned about heart health issues]. Abbott 2011 CSR rep. Necessity/obligation modal These suppliers must [comply with the expectations and standards of our code…]. Kellogg 2011 CSR rep. Volition/prediction modal We will [source all sunflower oil sustainably by 2020]. Unilever 2011 CSR rep. Stance type Construction type Illustrative corpus examplea Source Attitudinal stance Ability or willingness adjective + complement clauseb We are determined [to continue providing leadership within our sphere of influence]. Neslté 2011 CSR rep. Attitude adverb Amazingly, [covered bonds in Europe count as 100% liquid assets…]. JP Morgan 2011 annual rep. Attitude verb + complement clauseb,c I regret [that six people died working for Shell in 2011]. Shell 2011 annual rep. Desire/intention/decision verb + to-complement clause Merck wants [to offer its talented employees the opportunity to have an interesting career]. Merck 2011 CSR rep. Ease or difficulty adjective + to-complement claused It is inherently difficult [to predict the outcome of the regulatory proceedings involving our businesses]. HSBC 2011 annual rep. Emotion/attitude adjective + complement clauseb,c I am happy [to report that we also made major strides in those aspects of our strategy in 2011…]. Shell 2011 annual rep. Evaluation adjective + complement clauseb,c,d It is essential [that we meet consistently high ethical and quality standards for research and development in all parts of our business…]. GSK 2011 CSR rep. Epistemic stance Certainty/likelihood adjective + complement clauseb,c,d With stronger capital, liquidity and reserves, we are confident [that we will continue to succeed in the ever-changing global business environment]. Bank of America 2011 CSR rep. Certainty/likelihood adverb Real estate was not the only culprit in the recent crisis, but [it certainly was at the eye of the storm]. JP Morgan 2011 annual rep. Certainty/likelihood noun + that-complement clause It remains our strong belief [that operating in a responsible and ethical way is essential for the success of GSK]. GSK 2011 annual rep. Certainty/likelihood verb + complement clauseb,c Our analysis suggests [biofuels could make up as much as 23% of global incremental demand for transport fuels over the period 2010–2030]. BP 2011 CSR rep. Modality Permission/possibility/ability modal A bigger challenge may [be convincing women that they need to be concerned about heart health issues]. Abbott 2011 CSR rep. Necessity/obligation modal These suppliers must [comply with the expectations and standards of our code…]. Kellogg 2011 CSR rep. Volition/prediction modal We will [source all sunflower oil sustainably by 2020]. Unilever 2011 CSR rep. Note: aThe stance marker is highlighted in bold and the proposition over which it takes scope is enclosed within square brackets. b Only complement clauses occurring in post-predicate position were taken into account. All possible structural types of complement clause were considered, that is, that-clauses, wh-clauses, infinitive clauses, and ing-clauses (see Biber et al. 1999: Chapter 9). c Instances where the complementizer that was omitted were also considered (see Biber et al. 1999: 680–3). d Extraposed constructions were also considered (see Biber et al. 1999: 660). Table 1 provides a complete list of the stance constructions surveyed. Each construction is accompanied by an illustrative example taken from the corpus. 5 THE CORPUS For this study, a specialized corpus of roughly 2.5 million words was compiled. It contains the 2011 annual reports and CSR reports published by a sample of publicly traded multinational corporations belonging to four different industry sectors. The sectors considered are financial services, oil and gas, pharmaceuticals, and food processing. The corpus includes the 2011 annual and CSR reports published by the fourth largest Europe- or US-based companies2 in each of these sectors (i) that released both an annual report and a stand-alone CSR report for the year considered (2011), and (ii) whose annual report contains both legally required financial data (e.g. audited financial statements) and voluntary information (e.g. letter to shareholders, operating, and financial review). The choice of sectors was guided by the aim to optimize corpus diversity and representativeness within the resources available for this study; the sectors examined include companies with different stakeholder priorities and relations, and with different strategic, economic, and CSR challenges. Table 2 provides summary information about the corpus. As shown in the table, the corpus is subdivided into two subcorpora for comparative purposes: the annual report subcorpus (henceforth, ARC) and the CSR report subcorpus (henceforth, CSR). A detailed overview of the corpus, including the word count for each text, can be found in the Supplementary Materials. Table 2: Overview of the corpus Total number of reports 32 (16 annual reports; 16 CSR reports) Total corpus size 2,433,956 Size of subcorpora ACR 1,828,029 CSR 605,927 Mean report size ACR 114,251.8 CSR 37,870.4 Sectors and companies Financial services JP Morgan Chase HSBC Wells Fargo Bank of America Oil and gas Exxon Mobil Royal Dutch Shell Chevron BP Pharmaceuticals Johnson & Johnson Merck GlaxoSmithKline Abbott Laboratories Food processing Neslté Unilever General Mills Kellogg Total number of reports 32 (16 annual reports; 16 CSR reports) Total corpus size 2,433,956 Size of subcorpora ACR 1,828,029 CSR 605,927 Mean report size ACR 114,251.8 CSR 37,870.4 Sectors and companies Financial services JP Morgan Chase HSBC Wells Fargo Bank of America Oil and gas Exxon Mobil Royal Dutch Shell Chevron BP Pharmaceuticals Johnson & Johnson Merck GlaxoSmithKline Abbott Laboratories Food processing Neslté Unilever General Mills Kellogg Table 2: Overview of the corpus Total number of reports 32 (16 annual reports; 16 CSR reports) Total corpus size 2,433,956 Size of subcorpora ACR 1,828,029 CSR 605,927 Mean report size ACR 114,251.8 CSR 37,870.4 Sectors and companies Financial services JP Morgan Chase HSBC Wells Fargo Bank of America Oil and gas Exxon Mobil Royal Dutch Shell Chevron BP Pharmaceuticals Johnson & Johnson Merck GlaxoSmithKline Abbott Laboratories Food processing Neslté Unilever General Mills Kellogg Total number of reports 32 (16 annual reports; 16 CSR reports) Total corpus size 2,433,956 Size of subcorpora ACR 1,828,029 CSR 605,927 Mean report size ACR 114,251.8 CSR 37,870.4 Sectors and companies Financial services JP Morgan Chase HSBC Wells Fargo Bank of America Oil and gas Exxon Mobil Royal Dutch Shell Chevron BP Pharmaceuticals Johnson & Johnson Merck GlaxoSmithKline Abbott Laboratories Food processing Neslté Unilever General Mills Kellogg 5.1 Corpus compilation and text conversion The corpus compilation process involved three main steps. First, the pdf reports were downloaded from the respective companies’ Web sites. A complete list of the URLs from which the reports were downloaded is provided in the Supplementary Materials. Second, the pdf reports were automatically converted to raw text with the aid of the open-source computer program PDFMiner.3 Third, each corpus file was manually inspected to identify and correct conversion errors, such as misspelled words due to incorrect decoding of text characters, interrupted sentences due to column or page breaks, misspelled hyphenated words due to incorrect decoding of line breaks. The inspection and correction of the files were carried out in a step-wise fashion with the aid of the advanced text editor TextWrangler,4 using regular expressions to systematically identify and amend the errors. A detailed description of the text inspection and cleaning procedure is provided in the Supplementary Materials, including the regular expressions used at each step of the process. 6 METHOD The procedure for the identification and quantification of stance constructions in the corpus relied on a combination of automatic techniques and manual analysis. It comprised three steps: 1. Automatic stance marker retrieval. For each of the categories included in the analysis framework, an extended list of stance markers was created based on the lists provided in Biber and Finegan (1989) and Biber (2006b). These lists were integrated with additional stance markers derived from the author’s prior work, from an informal exploratory analysis of the corpus, and by consulting a thesaurus of English. By means of a string search Perl script,5 the stance markers included in the lists were simultaneously and automatically searched for in the corpus. For each marker, all inflected forms, for instance believe, believes, believed, believing, were included in the query. The script output consisted of a tabled list of the markers from the original set that were actually found in the corpus and the frequency of each. In this way, the expressions that were not used in the texts under study could immediately be identified and discarded, simplifying the following analysis steps. The complete lists of stance markers used at this stage are provided in Table A1 (available in the Supplementary Materials).6 The frequency lists produced by the string search script are given in the Supplementary Materials. 2. Manual word-sense disambiguation by means of concordance analysis. Some of the stance expressions detected in step 1 are polysemous and may carry both stance and non-stance senses. Furthermore, they may be used as lexical, rather than grammatical, stance markers, and may be attributed to third parties. Compare, for example: (7) (a) ExxonMobil is proud to be a charter member of the Wildlife Habitat Council (WHC) since its establishment in 1988. (Exxon CSR rep.)   (b) Whatever the outcome, Adeline’s sights are set firmly on qualifying for the 2016 Olympics in Rio de Janeiro, Brazil. She explains: ‘I want to do well for my SOS family, my SOS mother and my country. They helped me to get where I am now and I want them to be proud of me.’ (HSBC CSR rep.) In example (7a) proud is used as a grammatical marker of attitudinal stance in an emotion/attitude adjective + to-complement clause construction. Conversely, in example (7b), the same word is used as a lexical marker of stance to describe a desired state of affairs, rather than to express the speaker’s attitude toward a proposition. In addition, this sentence is attributed to a person that does not belong to the company, namely, one of the beneficiaries of a community development plan launched by HSBC bank. For these reasons, (7b) should be excluded from the analysis (cf. analysis criteria in Section 4.1). Since the string search algorithm used in step one can only retrieve single words, it does not consider the linguistic context in which the stance expressions are found, nor does it incorporate any automatic word-sense disambiguation component, it may produce a high number of false positives and return unreliable results. Therefore, the second step of the analysis consisted in manually inspecting the results of the automatic search to identify and remove all irrelevant items. This was accomplished by means of concordance analysis with the aid of the corpus program AntConc (Anthony 2012). A Key Word in Context (KWIC) concordance was produced for each of the words retrieved by the search string algorithm in step 1, and exported to a spreadsheet.7 Each concordance line was manually inspected and annotated as either relevant or irrelevant based on the criteria outlined in Section 4.1. In a limited number of cases, ad hoc rules had to be formulated to deal with problematic expressions and complex co-textual dynamics. A full account of these ad hoc rules is included in the Supplementary Materials. 3. Quantification and statistical analysis. Once concordances for all the stance markers retrieved in step 1 were produced and all irrelevant entries were removed from them, the frequency of occurrence of each marker in the two subcorpora was calculated. Frequencies were normalized to allow for a comparison between the subcorpora, which are of different size. The aggregated frequency of stance markers was calculated and compared across the categories included in the analysis framework. Where relevant, pairwise log-likelihood tests (Rayson and Garside 2000) were performed to determine whether the differences observed in the frequency of stance markers between the ARC and CSR subcorpora were statistically significant, that is, not attributable to chance variation.8p-values equal to or lower than .05 were considered statistically significant. The next section summarizes the results of the analysis and discusses a number of examples from the corpus. 7 RESULTS This section presents the results of the analysis. It begins by examining the overall distribution of stance constructions in the corpus, and then proceeds to discussing the main findings for each of the three general stance categories considered, that is, attitudinal stance, epistemic stance, and modality, in more detail. Due to space constraints, the presentation of the results will focus on the most important findings. The complete quantitative results are provided in the Supplementary Materials, including the raw and normalized frequencies for each stance marker and category surveyed. 7.1 General distribution of stance constructions Figure 1 shows the overall frequency of stance constructions in the ARC and CSR subcorpora, grouped by type. The plot shows that modals are by far the most common type of stance marker. Modals are significantly more frequent in the CSR compared to the ARC (LL = 11.96, df = 1, p < .001). Attitudinal stance constructions are also significantly more frequent in the CSR than in the ARC (LL = 945.31, df = 1, p<.001). A minor, non-significant difference was found in the frequency of epistemic stance markers between the two subcorpora (LL = 0.45, df = 1, p = .50). Overall, the CSR contains a significantly higher number of stance markers in comparison to the ARC—8674.22 versus 6801.22 instances per million words (LL = 209.58, df = 1, p < .001). Figure 1: View largeDownload slide Distribution of stance types across corpora Figure 1: View largeDownload slide Distribution of stance types across corpora From these findings, we may infer that the language in CSR reports is more explicitly subjective and evaluative than in annual reports. In the former, companies are more prone to openly express their emotions, attitudes, and assessments, as evidenced by the comparatively higher frequency of stance markers in general, and by the overwhelmingly higher frequency of attitudinal expressions in particular. To gain a deeper understanding of these differences and of how stance markers are used in annual and CSR reports, however, we need to take a closer look at each stance category. The next section provides a more detailed account of the distribution and functions of attitudinal stance constructions in the corpus. 7.2 Attitudinal stance Figure 2 reports the frequency and distribution of attitudinal stance constructions in the corpus. As the plot shows, desire/intention/decision verb + to-clause and ability or willingness adjective + complement clause are the two most frequent attitudinal stance constructions among the seven surveyed. Both are more than three times as frequent in the CSR compared to the ARC. Figure 2: View largeDownload slide Distribution of attitudinal stance constructions Figure 2: View largeDownload slide Distribution of attitudinal stance constructions The most common desire/intention/decision verbs controlling a complement clause are reported in Figure 3. The graph shows that aim is the most frequent verb in this group, and is overwhelmingly more frequent in the CSR compared to the ARC. This result suggests that, in their CSR reports, companies place considerable emphasis on objectives and ambitions for the future. These relate primarily to positive social and environmental goals (87.93% of all instances of aim + to-clause), as illustrated in example (8). (8) We aim to be a good neighbour to the communities close to our projects and facilities. (Shell CSR rep.) Statements of good intentions like the one in (8) often appear to be used rhetorically in CSR reports, to demonstrate the reporting company’s commitment to the principles of CSR, and project integrity. As shown in Figure 3, other common desire/intention/decision verbs in the CSR are seek, strive, and want. Similar to aim, these verbs are often used in CSR reports to emphasize a company’s commitment to ethical and responsible conduct (88.43% of all instances), and thus to communicate integrity, as illustrated in (9). (9) We strive to maintain the highest standards of integrity and do the right thing for our customers, colleagues, shareholders, and all of our other partners. (JP Morgan CSR rep.) Figure 3: View largeDownload slide Distribution of desire/intention/decision verbs Figure 3: View largeDownload slide Distribution of desire/intention/decision verbs While all three verbs are markedly less frequent in the ARC compared to the CSR, strive stands out as particularly rare in the former subcorpus. One possible explanation for this considerable difference resides in the semantics of this verb. Strive expresses a genuine and earnest effort toward a goal. It evokes positive evaluative meanings of determination, tenacity, and humbleness. However, striving for a goal does not necessarily imply that it will be achieved. Such a non-binding declaration of intents, even if loaded with positive valence, may not be sufficient to reassure shareholders and investors that the company will live up to their expectations and turn their investments into profit. The only desire/intention/decision verb that is comparatively more frequent in annual reports is intend. In 59.41% of the occurrences, this verb is used to discuss a company’s strategic and financial plans, as shown in (10). (10) Within the Refining and Marketing segment, BP intends to divest the Texas City refinery and related assets […]. (BP annual rep.) Compared to most other verbs in this group, intend is a relatively neutral stance verb. It communicates rationality, and indexes emotionally controlled decision-making. The relatively frequent use of this verb to communicate intentions in annual reports may be interpreted as reflecting the companies’ need to show shareholders and investors that they are in control of the situation, and that their decisions are rational, rather than emotional. Further, compared to other verbs in this category, intend expresses a more definite commitment to a goal, and may thus contribute to strengthen the impression that the company can be trusted to deliver on its targets. Ability or willingness adjective + complement clause is the second most frequent attitudinal stance construction in both the subcorpora and, as noted above, it is more frequent in the CSR than in the ARC. Overall, the most frequent ability/willingness adjectives are committed, able, and dedicated. In the vast majority of cases (91.67%), the words committed and dedicated are employed in CSR reports to foreground a company’s determination in pursuing social and environmental objectives, as in (11). (11) We are dedicated to playing our part in addressing this important public health challenge, and micronutrient fortification will remain a focus in 2012. (Nestlé CSR rep.) The expression committed to is overwhelmingly more frequent in the CSR than in the ARC—308.61 compared to 29.54 occurrences per million words (LL = 294.51, df = 1, p < .001). This difference accounts for a substantial part of the difference in the frequency of ability/willingness adjective + complement clause constructions across the subcorpora. Example (12) illustrates this construction. (12) At Abbott, we are committed to meeting high ethical standards and to complying with all applicable local, national, and international laws wherever we do business. (Abbott CSR rep.) From the point of view of corporate identity, the large difference observed in the frequency of committed to constructions between the subcorpora is highly significant. Semantically, the adjective committed expresses a strong degree of willingness, and carries meanings of obligation, loyalty, and accountability. Thus, the frequent use of this word in CSR reports can be seen to reinforce the idea that the reporting companies are strongly and genuinely determined to become better corporate citizens. At the same time, however, committed is similar to strive in that it signals a firm intention to reach a goal, but leaves the possibility open that the goal might not actually be achieved. This might explain why it is not as frequently used in annual reports as in CSR reports. Among ability or willingness adjective + complement clause constructions, able + to-clause is also relatively frequent. This construction is often used in both report types to underline a company’s achievements and capabilities (56.76% of all instances), as example (13) shows. (13) Through year-end 2011, we have been able to achieve significant emission reductions […]. (JP Morgan CSR rep.) Once again, this construction is significantly more frequent in CSR reports compared to annual reports, with 95.72 and 49.23 occurrences per million words, respectively (LL = 14.63, df = 1, p < .001). 7.3 Epistemic stance Figure 4 plots the distribution of epistemic stance constructions in the corpus. As the graph shows, certainty/likelihood verbs controlling a complement clause are by far the most frequent resources used to express epistemic stance, followed by adverbs, adjectives controlling a complement clause, and nouns controlling a complement clause. Figure 4: View largeDownload slide Distribution of epistemic stance constructions Figure 4: View largeDownload slide Distribution of epistemic stance constructions Following Biber and Finegan (1989) and Biber (2006b), stance constructions can be broadly classified into those that express certainty about a proposition, and those that express doubt or likelihood. Figure 5 represents the distribution of epistemic stance constructions into these two subcategories in the ARC and CSR. Constructions expressing certainty are significantly more frequent in the CSR (LL = 100.03, df = 1, p < .001), whereas markers of likelihood are significantly more common in the ARC (LL = 48.53, df = 1, p < .001). This finding indicates that companies are generally more assertive in CSR reports, whereas they tend to adopt a more cautious, tentative stance in annual reports. Figure 5: View largeDownload slide Distribution of certainty and likelihood constructions Figure 5: View largeDownload slide Distribution of certainty and likelihood constructions A closer look at the frequency of the epistemic verbs retrieved in the subcorpora offers further insights into the use of certainty and likelihood stance expressions in the corpus, and the communicative functions they perform. As Figure 6 shows, only four certainty verbs are more frequent in the ARC than in the CSR: anticipate, conclude, confirm, and determine. These verbs point to two major communicative functions performed by annual reports, that is, (i) evaluating the performance and the financial and legal standing of a company, and (ii) anticipating future scenarios, decisions, and actions. Examples (14) and (15) illustrate these two functions in turn. (14) The review concluded the board had operated well in 2011. (BP annual rep.) (15) The Company anticipates that operating cash flows, existing credit facilities and access to the commercial paper markets will provide sufficient resources to fund operating needs in 2012. (Johnson & Johnson annual rep.) Figure 6: View largeDownload slide Distribution of certainty verbs controlling a complement clause Figure 6: View largeDownload slide Distribution of certainty verbs controlling a complement clause The relatively frequent use of these verbs in the ARC has important implications for the discursive construction of corporate identity. By providing informed assessments of the present situation and by carefully considering and anticipating future scenarios, companies not only address shareholders’ information needs, but also project readiness and competence; they show that they are in control of the situation and prepared for any contingency. As far as the CSR is concerned, the most common certainty verbs are show, know, find, and understand. The verbs show and find serve three main communicative functions in CSR reports. First, they are used to provide empirical evidence for the companies’ accomplishments related to different aspects of CSR (29.35% of all instances). In (16), for example, Bank of America showcases its progress toward reduced water consumption. (16) In 2011, we completed a two-year comprehensive inventory of our water consumption (our first ever). The review found that we consumed 3.9 billion gallons of water in 2011, down from 4.2 billion gallons in 2010. (Bank of America CSR rep.) Second, these verbs are used to present scientifically credible evidence for the companies’ positive contribution to the solution of pressing social and environmental problems (26.09% of all instances). In (17), for instance, the cereal company General Mills refers to research demonstrating the potential positive effects of cereals against obesity, thus suggesting that the company’s business is inherently good. (17) A 2009 study showed that regardless of sweetness level, cereal eaters have healthier body weights than those who don’t eat cereal. (General Mills CSR rep.) Third, show and find are used in CSR reports to refer to scientific research that reportedly informs the companies’ decisions and strategies (20.65% of all instances). The verbs know and understand serve an important dialogic and solidarity-building function in CSR reports; they are often used to acknowledge the stakeholders’ perspective, and to show that the reporting company is sympathetic to their concerns (63.01% of all instances). This function is exemplified in (18), where GlaxoSmithKline openly recognizes the ethical implications of its activities, thereby implicitly suggesting that the company is aware of and willing to address them. (18) We know that the research and development, manufacture and sale of our products can raise ethical issues. (GSK CSR rep.) From the point of view of corporate identity, the repeated use of know and understand as perspective taking predicates, as illustrated in (18), serves to project benevolence. By explicitly acknowledging the risks and challenges inherent in their business activities and by demonstrating sympathy and a readiness to listen to the stakeholders’ views, companies present themselves as receptive and caring organizations (cf. Livesey and Kearins 2002). Figure 7 represents the distribution of likelihood verbs across the subcorpora. The plot reveals two outstanding differences between the ARC and the CSR; the verb expect is comparatively more frequent in the ARC, whereas believe is more frequent in the CSR. These differences are statistically significant (expect: LL = 112.14, df = 1, p < .001; believe: LL = 5.23, df = 1, p = .022). The frequent use of expect in the ARC mirrors the similarly frequent use of anticipate observed above. From a corporate identity perspective, the use of these anticipatory verbs can be seen to contribute to highlight the company’s readiness for the future. Compared to anticipate, however, expect conveys a more tentative stance, and serves to hedge forecasts. Therefore, from the company’s perspective, this verb can have the twofold beneficial effect of conveying a cautious and responsible attitude toward predictions, and of cushioning the negative effects of potential misjudgments. Example (19) is an example of this verb from the corpus. (19) We are well placed in emerging markets and we expect these markets to continue to drive growth. (Unilever annual rep.) Figure 7: View largeDownload slide Distribution of likelihood verbs Figure 7: View largeDownload slide Distribution of likelihood verbs Believe is the most frequent likelihood verb in CSR reports. It is often used to foreground the reporting company’s ideological commitment to the principles of CSR (35.94% of all instances), as in (20). (20) At JPMorgan Chase, we believe our place is right at the heart of efforts to solve daunting problems like unemployment, environmental sustainability, and access to quality education. (JP Morgan CSR rep.) As Fetzer (2008) observes, believe expresses a stronger degree of epistemic commitment compared to other likelihood verbs such as think, in particular when it is preceded by a first-person pronoun, as in the example above. The frequent use of this verb to communicate the companies’ ethical values and credo thus serves to emphasize their genuine and strong commitment to the cause of CSR, and to project integrity. Believe is a very important feature of the language of annual reports as well, being the second most frequent likelihood verb. It performs three primary communicative functions in these texts. First, it is used to present management’s assessment of the company’s performance and financial situation, and of the risks it faces. This function accounts for 41.26% of the occurrences, and is illustrated in example (21). (21) We believe that our portfolio of assets remains well positioned to compete and grow value in a range of external conditions and we continue to increase both investment and operating cash. (BP annual rep.) Second, believe is frequently used to explain and justify management’s operational and strategic decisions (27.79% of all instances), as shown in (22). (22) We have focus. We do not compete in numerous categories as we believe expertise in a few businesses is preferable to having simple scale as the result of participation in many. (Kellogg annual rep.) Third, in 12.21% of the occurrences, believe is used to express opinions related to legal proceedings in which the company is involved, as shown in (23). (23) Based on the history described above, Chevron believes that this lawsuit lacks legal or factual merit. (Chevron annual rep.) A substantial proportion of the instances of believe found in annual reports occurs in the introductory message to shareholders (13.26%), where the CEO and/or Chairman review the company’s achievements, and summarize goals and expectations. In this context, believe is generally used to highlight the company’s strengths, and foster optimism about the future, as example (24) shows. (24) I am proud to be on this team and believe that our best days are ahead. (JP Morgan annual rep.) In all the cases reviewed above, believe appears to be used to promote trust in the company’s competence and ability to overcome difficulties and succeed. 7.4 Modality Figure 8 displays the distribution of modals across the three main categories in the ARC and CSR. Obligation/necessity modals are significantly more frequent in the CSR (LL = 39.43, df = 1, p < .001), permission/possibility/ability modals are significantly more common in the ARC (LL = 20.02, df = 1, p < .001), and volition/prediction modals occur significantly more often in the CSR (LL = 49.38, df = 1, p < .001). Figure 8: View largeDownload slide Distribution of modal constructions Figure 8: View largeDownload slide Distribution of modal constructions Figure 9 shows the frequency of permission/possibility/ability modals in the subcorpora. Two major differences between the ARC and the CSR can be observed: can is more than twice as frequent in the CSR than the ARC; may and could are both substantially more frequent in the ARC compared to the CSR. Figure 9: View largeDownload slide Distribution of permission/possibility/ability modals Figure 9: View largeDownload slide Distribution of permission/possibility/ability modals The comparatively higher frequency of the modals may and could in the ARC is in line with the pattern observed above for likelihood verbs, namely, that companies tend to adopt a more tentative stance in annual reports. But the relatively higher frequency of these modals also reflects the fact that one of the chief communicative functions of annual reports is that of discussing and anticipating potential, future scenarios. Not surprisingly, among the most frequent verbs appearing in the immediate-right collocate slot for may in the ARC, which are shown in Table 3, we find result, require, become, affect, vary, and change. These verbs can be related to the implications and consequences of anticipated events for the company and its investors. Table 3: Verbal collocates of may ARC CSR Collocate T-score Collocate T-score be 25.19 be 7.98 have 8.88 have 3.29 include 8.54 affect 2.98 result 7.90 differ 2.64 require 6.37 result 2.41 differ 6.12 include 2.36 become 5.93 contact 2.21 affect 5.53 exist 1.99 vary 5.52 occur 1.99 change 4.92 lead 1.96 ARC CSR Collocate T-score Collocate T-score be 25.19 be 7.98 have 8.88 have 3.29 include 8.54 affect 2.98 result 7.90 differ 2.64 require 6.37 result 2.41 differ 6.12 include 2.36 become 5.93 contact 2.21 affect 5.53 exist 1.99 vary 5.52 occur 1.99 change 4.92 lead 1.96 Table 3: Verbal collocates of may ARC CSR Collocate T-score Collocate T-score be 25.19 be 7.98 have 8.88 have 3.29 include 8.54 affect 2.98 result 7.90 differ 2.64 require 6.37 result 2.41 differ 6.12 include 2.36 become 5.93 contact 2.21 affect 5.53 exist 1.99 vary 5.52 occur 1.99 change 4.92 lead 1.96 ARC CSR Collocate T-score Collocate T-score be 25.19 be 7.98 have 8.88 have 3.29 include 8.54 affect 2.98 result 7.90 differ 2.64 require 6.37 result 2.41 differ 6.12 include 2.36 become 5.93 contact 2.21 affect 5.53 exist 1.99 vary 5.52 occur 1.99 change 4.92 lead 1.96 Compared to may and could, can expresses a more assertive stance. The comparatively higher frequency of can in CSR reports appears to be coherent with the relatively higher frequency of certainty verbs discussed above. An inspection of the immediate-right verbal collocates of this modal verb, listed in Table 4, reveals that can is frequently accompanied in CSR reports by positively valenced verbs such as help, play (primarily used in the phrase play a role), and contribute. These expressions are used to underline the potential positive outcomes of a company’s decisions and actions, as in (25), or to discuss the potential benefits of responsible business practices, as in (26). Table 4: Verbal collocates of can ARC CSR Collocate T-score Collocate T-score be 21.99 be 12.86 vary 3.97 help 6.41 do 3.85 make 4.97 affect 3.79 have 4.89 have 3.41 read 4.86 help 3.41 do 4.09 make 3.19 play 3.83 provide 3.14 contribute 3.83 take 2.92 take 3.53 include 2.84 provide 3.35 ARC CSR Collocate T-score Collocate T-score be 21.99 be 12.86 vary 3.97 help 6.41 do 3.85 make 4.97 affect 3.79 have 4.89 have 3.41 read 4.86 help 3.41 do 4.09 make 3.19 play 3.83 provide 3.14 contribute 3.83 take 2.92 take 3.53 include 2.84 provide 3.35 Table 4: Verbal collocates of can ARC CSR Collocate T-score Collocate T-score be 21.99 be 12.86 vary 3.97 help 6.41 do 3.85 make 4.97 affect 3.79 have 4.89 have 3.41 read 4.86 help 3.41 do 4.09 make 3.19 play 3.83 provide 3.14 contribute 3.83 take 2.92 take 3.53 include 2.84 provide 3.35 ARC CSR Collocate T-score Collocate T-score be 21.99 be 12.86 vary 3.97 help 6.41 do 3.85 make 4.97 affect 3.79 have 4.89 have 3.41 read 4.86 help 3.41 do 4.09 make 3.19 play 3.83 provide 3.14 contribute 3.83 take 2.92 take 3.53 include 2.84 provide 3.35 (25) Achieving higher sales volumes can help to reduce the costs of goods which we can pass on as further price reductions […]. (GSK CSR rep.) (26) Sustainable sourcing can contribute to better incomes and livelihoods for farmers and their workers. (Unilever CSR rep.) As far as volition/prediction modals are concerned, will is comparatively more frequent in the CSR, whereas would is more frequent in the ARC, as shown in Figure 10. The modal shall is infrequent in both subcorpora. Both will and would can be used to make forecasts about the future. However, compared to would, will conveys a more assertive, confident stance. The difference in the distribution of will and would between the subcorpora seems to confirm the patterns observed for the modals may and can, namely that companies tend to show a more cautious attitude to forecasts in annual reports than in CSR reports. Figure 10: View largeDownload slide Distribution of volition/prediction modals Figure 10: View largeDownload slide Distribution of volition/prediction modals To conclude the survey of modal markers, I now turn to necessity modals. As previously noted, necessity modals are overall more frequent in the CSR. Figure 11 reveals that the most striking difference between the subcorpora concerns must. An analysis of the immediate-right verbal collocates of this modal verb, provided in Table 5, shows that it is frequently accompanied in CSR reports by verbs such as follow, adhere, comply, which are commonly used with reference to the rules, guidelines, and standards that a company’s employees, contractors, and suppliers must observe (91.67% of all instances), as illustrated in (27). (27) All employees are expected to uphold the highest ethical standards of business integrity. They must comply with all applicable laws and accurately record and track all business transactions. (Unilever CSR rep.) Table 5: Verbal collocates of must ARC CSR Collocate T-score Collocate T-score be 10.16 be 7.01 have 3.42 meet 2.60 meet 2.42 follow 2.44 prove 1.99 provide 2.15 develop 1.98 continue 1.93 make 1.95 work 1.81 provide 1.93 adhere 1.72 maintain 1.71 complete 1.72 manage 1.69 comply 1.71 state 1.68 create 1.69 ARC CSR Collocate T-score Collocate T-score be 10.16 be 7.01 have 3.42 meet 2.60 meet 2.42 follow 2.44 prove 1.99 provide 2.15 develop 1.98 continue 1.93 make 1.95 work 1.81 provide 1.93 adhere 1.72 maintain 1.71 complete 1.72 manage 1.69 comply 1.71 state 1.68 create 1.69 Table 5: Verbal collocates of must ARC CSR Collocate T-score Collocate T-score be 10.16 be 7.01 have 3.42 meet 2.60 meet 2.42 follow 2.44 prove 1.99 provide 2.15 develop 1.98 continue 1.93 make 1.95 work 1.81 provide 1.93 adhere 1.72 maintain 1.71 complete 1.72 manage 1.69 comply 1.71 state 1.68 create 1.69 ARC CSR Collocate T-score Collocate T-score be 10.16 be 7.01 have 3.42 meet 2.60 meet 2.42 follow 2.44 prove 1.99 provide 2.15 develop 1.98 continue 1.93 make 1.95 work 1.81 provide 1.93 adhere 1.72 maintain 1.71 complete 1.72 manage 1.69 comply 1.71 state 1.68 create 1.69 Figure 11: View largeDownload slide Distribution of obligation/necessity modals Figure 11: View largeDownload slide Distribution of obligation/necessity modals The frequent use of necessity modals contributes to the discursive construction of the reporting companies’ responsible corporate identity; by means of these verbs, they give prominence to their ‘moral side’ and portray themselves as authoritative, responsible supervisors (cf. Fuoli 2012). 8 DISCUSSION The results of the analysis have revealed substantial differences in the way the companies included in the sample use stance expressions to construct their corporate identity in annual and CSR reports. Table 6 summarizes the main identity-building strategies identified in the two text types, alongside the textual evidence for them. The analysis has shown that, in annual reports, the companies examined attempt to present themselves as unbiased, rational, and competent decision makers. They display an objective and emotionally detached attitude toward the information provided, as evidenced by the relatively infrequent use of stance expressions in general, and of attitudinal markers in particular. They seek to convey rationality and resolve through the relatively frequent use of the verb intend. They show a cautious and responsible attitude to forecasts and assessments, as suggested by the frequent use of likelihood verbs and the modals may and could. They project competence by providing careful assessments of organizational performance and risks and reasoned justifications for their decisions through the verb believe, and demonstrate readiness by means of verbs such as expect and anticipate, which are recurrently used to discuss future scenarios. Table 6: Summary of main findings ARC CSR Identity-building strategy Textual evidence Identity-building strategy Textual evidence Display emotionally detached, objective attitude Relatively infrequent attitudinal stance expressions Highlight the ‘human side’ of the corporation Frequent attitudinal stance expressions Display rationality and definite commitment to goals Relatively frequent use of the verb intend Display firm and genuine commitment to corporate social responsibility; project integrity Frequent expressions of intentions and desires Display cautious and responsible attitude Frequent use of likelihood epistemic stance markers Display confident attitude in the face of potential skepticism Frequent use of certainty epistemic stance markers Project readiness and competence Frequent use of anticipate and expect to discuss future scenarios; frequent use of likelihood verb believe to carefully assess organizational strengths and risks, and to justify operational and strategic decisions Display heartfelt commitment to CSR; project integrity Frequent use of likelihood verb believe to present the company’s values Display cautious and responsible attitude; project readiness Frequent use of modals may and could to discuss future scenarios Provide empirical evidence for commitment to CSR Frequent use of certainty verbs show and find Display receptive and sympathetic attitude; project benevolence Frequent use of certainty verbs know and understand as perspective-taking predicates Highlight company’s positive contribution to environmental and social progress; project integrity Frequent use of modal can accompanied be by positively connoted verbs Portray the company as attentive supervisor; project integrity Frequent use of modal must ARC CSR Identity-building strategy Textual evidence Identity-building strategy Textual evidence Display emotionally detached, objective attitude Relatively infrequent attitudinal stance expressions Highlight the ‘human side’ of the corporation Frequent attitudinal stance expressions Display rationality and definite commitment to goals Relatively frequent use of the verb intend Display firm and genuine commitment to corporate social responsibility; project integrity Frequent expressions of intentions and desires Display cautious and responsible attitude Frequent use of likelihood epistemic stance markers Display confident attitude in the face of potential skepticism Frequent use of certainty epistemic stance markers Project readiness and competence Frequent use of anticipate and expect to discuss future scenarios; frequent use of likelihood verb believe to carefully assess organizational strengths and risks, and to justify operational and strategic decisions Display heartfelt commitment to CSR; project integrity Frequent use of likelihood verb believe to present the company’s values Display cautious and responsible attitude; project readiness Frequent use of modals may and could to discuss future scenarios Provide empirical evidence for commitment to CSR Frequent use of certainty verbs show and find Display receptive and sympathetic attitude; project benevolence Frequent use of certainty verbs know and understand as perspective-taking predicates Highlight company’s positive contribution to environmental and social progress; project integrity Frequent use of modal can accompanied be by positively connoted verbs Portray the company as attentive supervisor; project integrity Frequent use of modal must Table 6: Summary of main findings ARC CSR Identity-building strategy Textual evidence Identity-building strategy Textual evidence Display emotionally detached, objective attitude Relatively infrequent attitudinal stance expressions Highlight the ‘human side’ of the corporation Frequent attitudinal stance expressions Display rationality and definite commitment to goals Relatively frequent use of the verb intend Display firm and genuine commitment to corporate social responsibility; project integrity Frequent expressions of intentions and desires Display cautious and responsible attitude Frequent use of likelihood epistemic stance markers Display confident attitude in the face of potential skepticism Frequent use of certainty epistemic stance markers Project readiness and competence Frequent use of anticipate and expect to discuss future scenarios; frequent use of likelihood verb believe to carefully assess organizational strengths and risks, and to justify operational and strategic decisions Display heartfelt commitment to CSR; project integrity Frequent use of likelihood verb believe to present the company’s values Display cautious and responsible attitude; project readiness Frequent use of modals may and could to discuss future scenarios Provide empirical evidence for commitment to CSR Frequent use of certainty verbs show and find Display receptive and sympathetic attitude; project benevolence Frequent use of certainty verbs know and understand as perspective-taking predicates Highlight company’s positive contribution to environmental and social progress; project integrity Frequent use of modal can accompanied be by positively connoted verbs Portray the company as attentive supervisor; project integrity Frequent use of modal must ARC CSR Identity-building strategy Textual evidence Identity-building strategy Textual evidence Display emotionally detached, objective attitude Relatively infrequent attitudinal stance expressions Highlight the ‘human side’ of the corporation Frequent attitudinal stance expressions Display rationality and definite commitment to goals Relatively frequent use of the verb intend Display firm and genuine commitment to corporate social responsibility; project integrity Frequent expressions of intentions and desires Display cautious and responsible attitude Frequent use of likelihood epistemic stance markers Display confident attitude in the face of potential skepticism Frequent use of certainty epistemic stance markers Project readiness and competence Frequent use of anticipate and expect to discuss future scenarios; frequent use of likelihood verb believe to carefully assess organizational strengths and risks, and to justify operational and strategic decisions Display heartfelt commitment to CSR; project integrity Frequent use of likelihood verb believe to present the company’s values Display cautious and responsible attitude; project readiness Frequent use of modals may and could to discuss future scenarios Provide empirical evidence for commitment to CSR Frequent use of certainty verbs show and find Display receptive and sympathetic attitude; project benevolence Frequent use of certainty verbs know and understand as perspective-taking predicates Highlight company’s positive contribution to environmental and social progress; project integrity Frequent use of modal can accompanied be by positively connoted verbs Portray the company as attentive supervisor; project integrity Frequent use of modal must By contrast, in CSR reports, the companies included in this study use stance resources to portray themselves as committed, honest, and caring corporate citizens. They underscore their genuine commitment to the principles of CSR through the frequent use of expressions of intentions and desires. They strive to project integrity by emphasizing their ethical values through the verb believe, by highlighting their positive contribution to social and environmental progress, and by portraying themselves as attentive supervisors through the frequent use of necessity modals in relation to the rules they expect coworkers to observe. They seek to communicate benevolence by displaying a receptive and sympathetic attitude through the frequent use of know and understand as perspective-taking predicates. They foreground their ‘human side’ by frequently expressing desires, emotions, and evaluations. Overall, then, the results of this study suggest that companies tend to profile different identities in annual and CSR reports. If we accept the premise that corporate identity can function as a rhetorical, image-building tool, we can interpret the identities expressed in these two text types as strategic self-representations aimed at maximizing the persuasive appeal of the reports vis-á-vis the specific readerships they target. As mentioned above, shareholders, the principal intended readers of annual reports, primarily seek to maximize profits and minimize financial risk (Johansen and Ellerup Nielsen 2011). Since the profitability of their investment in the company crucially depends on the management’s ability to make sound decisions, they need and expect the company to be competently and responsibly managed. The discursive identity conveyed in annual reports seems to successfully address these requirements. By representing themselves as unbiased, rational, and competent decision makers, companies effectively respond to shareholders’ expectations, thereby laying the foundations for trust. CSR reports, on the other hand, address a more diverse set of stakeholders, including both internal and external groups. External stakeholders have been shown to attach particular importance to companies’ benevolence and integrity (Pirson and Malhotra 2011). Similarly, one of the characteristics that internal stakeholders with a deep relationship with the organization, such as employees, value the most is benevolence (Pirson and Malhotra 2011). As in the case of annual reports, the corporate identity projected in CSR reports appears to effectively target the intended audiences’ specific expectations. Companies mainly emphasize their integrity and benevolence in these texts, rather than other aspects of their identity. In sum, based on the results of the analysis, we may conclude that companies discursively construct and communicate distinct identities in annual and CSR reports, which appear to be strategically designed to meet the expectations of the reports’ primary target audiences, and thus to maximize the trust-building force of the texts. The results of the analysis confirm and complement some observations made in previous research on the genres of annual and CSR reports. As far as annual reports are concerned, the findings of this study provide some evidence for the dual nature of these texts as both informative and promotional texts (Bhatia 2004; Garzone 2004). However, the persuasive work performed by annual reports appears to be rather subtle, at least in comparison to CSR reports. The lower frequency of stance markers, and in particular of attitudinal stance markers, indicates that comparatively less room is given to explicitly evaluative expressions in annual reports, and thus to direct attempts to frame the reading of the information given. CSR reports appear to be relatively more explicitly subjective and evaluative, as evidenced by the significantly higher frequency of stance markers, in particular of the attitudinal type. This difference would seem to suggest that, while both report types are ‘hybrid’ discourse genres, the promotional component is more prominent in CSR reports. The tendency observed here for companies to display a caring, benevolent identity in CSR reports had already been noted by Livesey and Kearins (2002), who analyzed the pioneering CSR reports published by Shell and The Body Shop International in 1998. This study provides additional and more extensive evidence for the importance of this discursive strategy. The results of this study also confirm the previous finding that companies place considerable emphasis on commitments and on laudable goals in CSR reports (Skulstad 2008; Catenaccio 2011; Bondi 2016). My conclusions are in line with Bondi’s (2016) interpretation of this discursive maneuver as an attempt to create a ‘consistent, caring, and collaborative image of the company’ (Bondi 2016: 75). From a genre perspective, the very high frequency of expressions of desires and intentions suggests that these are distinctive features of CSR reports, and lends support to the contention that ‘making promises’ is a central rhetorical move in the genre (Skulstad 2008; Catenaccio 2011). From a critical perspective, this finding appears to at least partially corroborate the observation made by several critical scholars that good intentions and declarations of principles are given more prominence than concrete actions and achievements in CSR reports (Laine 2005; Milne et al. 2005). Clearly, more work is needed to empirically test this hypothesis, but the patterns detected here do raise important questions about the credibility of CSR reporting and about companies’ accountability. As Bondi (2016: 76) puts it, ‘planning to “be good” can certainly be appreciated, but is also hardly disputed’. One of the main findings of the analysis is the significantly higher frequency of certainty markers in CSR reports compared to annual reports. At first sight, this pattern may appear counter-intuitive. Considering that one of the chief functions of annual reports is that of soliciting or retaining the financial support of investors—a vital resource for public companies—we may reasonably expect organizations to display more confidence in these texts compared to CSR reports, where, from a communicative point of view, the stakes are lower. However, annual reports are directed at a readership of experts, who expect companies to objectively report on their performance and achievements. Thus, the target readers may view an overly confident attitude with suspicion. CSR reports, on the other hand, address a potentially skeptical readership. Skepticism toward CSR stems from two main facts. First, corporate social performance is complex to measure, due to the lack of universally accepted definitions and assessment criteria (McWilliams et al. 2006; Turker 2009). The reliability and adequacy of the data reported in CSR reports are, therefore, potentially questionable. Second, CSR communication is considered by many as a mere self-serving PR exercise carried out by companies to ‘greenwash’ their image and reputation (Skarmeas and Leonidou 2013; Bachmann and Ingenhoff 2016). For these reasons, companies need to make an extra effort to convince the reader about the reliability of their report, and about the truthfulness of the facts and figures included in it. In this light, it is unsurprising that companies frequently use certainty verbs such as show and find to provide empirical evidence for the sincerity and strength of their commitment to CSR. The frequent use of these verbs may be interpreted as a reactive discursive strategy aimed at neutralizing public skepticism by counteracting the actual or potential negative discourses surrounding the company’s CSR activities (Fuoli and Paradis 2014). Similar to the other strategies discussed above, the use of this type of stance markers exposes the inherent dialogic nature of corporate communication, and the strategic efforts made by companies to anticipate the readers’ attitudes and reactions to manage their impressions and steer them in a favorable direction. 9 CONCLUSION This study has investigated the use of stance expressions in annual and CSR reports with the aim to shed light on the discursive strategies that companies deploy to shape their corporate identity and gain the trust of the stakeholder groups that these two genres target. The analysis has revealed significant differences in the frequency and communicative functions of stance expressions in annual and CSR reports. The findings support the view that companies profile different identities in these texts to meet the expectations of the reports’ primary target audiences and maximize the trust-building force of their communications. The results of this study contribute to shed some new light on the identity work performed by companies in their reports. They also provide novel and linguistically informed insights into the impression management tactics used by companies in annual and CSR reports, by showing the important role played by stance expressions in promoting a trustworthy corporate image. Further, the study increases our knowledge of the genres of annual and CSR reports by providing a systematic analysis of stance features, and by highlighting significant differences between these two core members of the family of business reporting genres. Finally, the study provides both linguists and business communication scholars with a robust descriptive basis for critically assessing financial and CSR reporting. In conclusion, some limitations of the study should be noted. First, the analysis is restricted to grammatical markers of stance. Future work might consider extending the analysis to include lexical markers, and possibly multimodal aspects as well. Second, due to space limitations, the distribution of stance markers within and across texts, companies and industries has not been considered. This is, however, a potentially important aspect, which would add greater depth and precision to the analysis. Third, the analysis has focused on a relatively limited set of companies and economic sectors. Future studies could take into account additional organizations and industries to gain a more comprehensive and representative picture of the stance expressions and of the image-building strategies used in annual and CSR reports. Finally, the manual procedure for disambiguating stance constructions adopted here is very time-consuming. Future studies should focus efforts on automatizing this process to expedite the analysis and improve the scalability of the method. Acknowledgements I wish to thank Carita Paradis (Lund University, Sweden) for her valuable advice and help in editing and proofreading the manuscript. I would also like to thank the Editorial Board of Applied Linguistics and three anonymous reviewers for their constructive and thought-provoking comments on earlier versions of the article. Conflict of interest statement. None declared. NOTES Footnotes 1 Throughout the article, I use the term ‘speaker’ to refer to both speakers and writers. 2 Based on the 2011 Financial Times Global 500 ranking. URL: http://im.ft-static.com/content/images/33558890-98d4-11e0-bd66-00144feab49a.pdf [Accessed 20 May 2016] 3 PDFMiner is available, free of charge, at this URL: http://www.unixuser.org/∼euske/python/pdfminer/ [Accessed 20 May 2016] 4 TextWrangler is available, free of charge, at this URL: http://www.barebones.com/products/textwrangler/ [Accessed 20 May 2016] 5 The script, which is called ‘filter_by_field.pl’, is included in the supplementary materials for the following article: M. Baroni and A. Lenci. Distributional Memory: A general framework for corpus-based semantics. 2010. Computational Linguistics 36(4): 673–721. It can be downloaded, free of charge, from this URL: http://clic.cimec.unitn.it/dm/materials/scripts.tar.gz [Accessed 20 May 2016]. Instructions on how to run the script are available on demand. 6 It should be noted that this analysis step is not strictly necessary. The analysis could start from step 2, with a manual search of all the words included in the initial list. However, given the high number of potential markers, the procedure followed in step 1 greatly simplified and expedited the analysis process. 7 The concordance lines spanned 200 words to the left and to the right of the search word. 8 The log-likelihood tests were performed using the wizard developed by Andrew Hardie (Lancaster University, UK), and available at this URL: http://corpora.lancs.ac.uk/sigtest/ [Accessed 20 May 2016]. The tests were performed on the raw frequencies. REFERENCES Anderson C. J. , Imperia G. . 1992 . ‘The corporate annual report: A photo analysis of male and female portrayals ,’ Journal of Business Communication 29 : 113 – 28 . Google Scholar Crossref Search ADS Anthony L. 2012 . Antconc (version 3.2.4) [computer software]. Waseda University. Available at http://www.antlab.sci.waseda.ac.jp/. Accessed 20 May 2016. Bachmann P. , Ingenhoff D. . 2016 . ‘Legitimacy through CSR disclosures? The advantage outweighs the disadvantages ,’ Public Relations Review 42 : 386 – 94 . Google Scholar Crossref Search ADS Beattie V. , Dhanani A. , Jones M. . 2008 . ‘Investigating presentational change in U.K. annual reports. A longitudinal perspective ,’ Journal of Business Communication 45 : 181 – 222 . Google Scholar Crossref Search ADS Bebbington J. , Larrinaga C. , Moneva J. . 2008 . ‘Corporate social reporting and reputation risk management ,’ Accounting, Auditing and Accountability Journal 21 : 337 – 61 . Google Scholar Crossref Search ADS Bednarek M. 2015 . ‘“What we contrarians already know”: Individual and communal aspects of attitudinal identity’ in Groom N. , Charles M. , Suganthi J. (eds): Corpora, Grammar and Discourse: In Honour of Susan Hunston . John Benjamins , pp. 257 – 82 . Bhatia V. 2004 . Worlds of Written Discourse: A Genre-based View . Continuum . Biber D. 2004 . ‘Historical patterns for the grammatical marking of stance: A cross-register comparison ,’ Journal of Historical Pragmatics 5 : 107 – 36 . Google Scholar Crossref Search ADS Biber D. 2006a . ‘Stance in spoken and written university registers ,’ Journal of English for Academic Purposes 5 : 97 – 116 . Google Scholar Crossref Search ADS Biber D. 2006b . University Language: A Corpus-based Study of Spoken and Written Registers . John Benjamins . Biber D. , Finegan E. . 1988 . ‘Adverbial stance types in English ,’ Discourse Processes 11 : 1 – 34 . Google Scholar Crossref Search ADS Biber D. , Finegan E. . 1989 . ‘Styles of stance in English: Lexical and grammatical marking of evidentiality and affect ,’ Text and Talk 9 : 93 – 124 . Biber D. , Johansson S. , Leech G. , Conrad S. , Finegan E. . 1999 . Longman Grammar of Spoken and Written English . Pearson Education Limited . Bondi M. 2016 . ‘The future in reports: Prediction, commitment and legitimization in CSR ,’ Pragmatics and Society 7 : 57 – 81 . Google Scholar Crossref Search ADS Breeze R. 2012 . ‘Legitimation in corporate discourse: Oil corporations after Deepwater Horizon ,’ Discourse and Society 23 : 3 – 18 . Google Scholar Crossref Search ADS Breeze R. 2013 . Corporate Discourse . Bloomsbury . Buhr N. , Reiter S. . 2006 . ‘Ideology, the environment and one world view: A discourse analysis of Noranda’s environmental and sustainable development reports’ in Freedman M. , Jaggi B. (eds): Advances in Environmental Accounting & Management , vol. 3 . Emerald Group Publishing Limited , pp. 1 – 48 . Carroll A. B. 1999 . ‘Corporate social responsibility. Evolution of a definitional construct ,’ Business and Society 38 : 268 – 95 . Google Scholar Crossref Search ADS Catenaccio P. 2011 . ‘Social and environmental reports: A diachronic perspective on an emerging genre’ in Grazone G. , Gotti M. (eds): Discourse, Communication and the Enterprise: Genres and Trends . Peter Lang , pp. 169 – 91 . Chafe W. , Nichols J. . 1986 . Evidentiality: The Linguistic Coding of Epistemology . Ablex Publishing Corporation . Conrad S. , Biber D. . 2000 . ‘Adverbial marking of stance in speech and writing’ in Hunston S. , Thompson G. (eds): Evaluation in Text: Authorial Stance and the Construction of Discourse . Oxford University Press , pp. 56 – 73 . Dahlsrud A. 2006 . ‘How corporate social responsibility is defined: An analysis of 37 definitions ,’ Corporate Social Responsibility and Environmental Management 15 : 1 – 13 . Google Scholar Crossref Search ADS David C. 2001 . ‘Mythmaking in annual reports ,’ Journal of Business and Technical Communication 15 : 195 – 222 . Google Scholar Crossref Search ADS Deegan C. 2014 . ‘An overview of legitimacy theory as applied within the social and environmental accounting literature’ in Bebbington J. , Unerman J. , O’Dweyer B. (eds): Sustainability Accounting and Accountability , 2 nd edn Routledge , pp. 248 – 72 . Ditlev-Simonsen C. D. , Wenstøp S. . 2012 . ‘Companies’ ethical commitment: An analysis of the rhetoric in CSR reports ,’ Issues in Social and Environmental Accounting 5 : 65 – 81 . Google Scholar Crossref Search ADS Ditlevsen M. G. 2012a . ‘Revealing corporate identities in annual reports ,’ Corporate Communications: An International Journal 17 : 379 – 403 . Google Scholar Crossref Search ADS Ditlevsen M. G. 2012b . ‘Telling the story of Danisco’s annual reports (1935 through 2007-2008) from a communicative perspective ,’ Journal of Business and Technical Communication 26 : 92 – 115 . Google Scholar Crossref Search ADS Dowling J. , Pfeffer J. . 1975 . ‘Organizational legitimacy: Social values and organizational behavior ,’ Pacific Sociological Review 18 : 122 – 36 . Google Scholar Crossref Search ADS Englebretson R. 2007 . Stancetaking in Discourse: Subjectivity, Evaluation, Interaction . John Benjamins . European Commission . 2001 . Green paper. Promoting a European framework for corporate social responsibility. Available at http://europa.eu/rapid/press-release_DOC-01-9_en.pdf. Accessed 20 May 2016. Favotto A. , Kollman K. , Bernhagen P. . 2016 . ‘Engaging firms: The global organisational field for corporate social responsibility and national varieties of capitalism ,’ Policy and Society 35 : 13 – 27 . Google Scholar Crossref Search ADS Fetzer A. 2008 . ‘“And I think that is a very straightforward way of dealing with it”: The communicative function of cognitive verbs in political discourse ,’ Journal of Language and Social Psychology 27 : 384 – 96 . Google Scholar Crossref Search ADS Filimon A. I. 2009 . ‘Kyosei—an example of cultural keyword argumentatively exploited in corporate reporting discourse ,’ Studies in Communication Sciences 9 : 131 – 52 . Fombrun C. 1996 . Reputation: Realizing Value from the Corporate Image . Harvard Business School Press . Freeman R. E. 1984 . Strategic Management: A Stakeholder Approach . Pitman . Friedman A. L. , Miles S. . 2006 . Stakeholders: Theory and Practice . Oxford University Press . Fuoli M. 2012 . ‘Assessing social responsibility: A quantitative analysis of Appraisal in BP’s and IKEA’s social reports ,’ Discourse and Communication 6 : 55 – 81 . Google Scholar Crossref Search ADS Fuoli M. , Hommerberg C. . 2015 . ‘Optimizing transparency, reliability and replicability: Annotation principles and inter-coder agreement in the quantification of evaluative expressions ,’ Corpora 10 : 315 – 49 . Google Scholar Crossref Search ADS Fuoli M. , Paradis C. . 2014 . ‘A model of trust-repair discourse ,’ Journal of Pragmatics 74 : 52 – 69 . Google Scholar Crossref Search ADS Garzone G. 2004 . ‘Annual company reports and CEO’s letters: Discoursal features and cultural markedness’ in Candlin C. , Gotti M. (eds): Intercultural Aspects of Specialized Communication . Peter Lang , pp. 311 – 43 . Garzone G. 2005 . ‘Letters to shareholders and chairman’s statements: Textual variability and generic integrity’ in Gillaerts P. , Gotti M. (eds): Genre Variation in Business Letters . Peter Lang , pp. 179 – 204 . Goletani L. 2011 . ‘Gazprom’s environmental report: Peculiarities of an emerging genre’, in Grazone G. , Gotti M. (eds): Discourse, Communication and the Enterprise: Genres and Trends . Peter Lang , pp. 255 – 277 . Gray R. , Javad M. , Power D. , Sinclair C. . 2001 . ‘Social and environmental disclosure and corporate characteristics: A research note and extension ,’ Journal of Business Finance and Accounting 28 : 327 – 56 . Google Scholar Crossref Search ADS Gray R. , Owen D. , Adams C. . 1996 . Accounting and Accountability: Changes and Challenges in Corporate Social and Environmental Reporting . Prentice Hall . Hallahan K. , Holtzhausen D. , Van Ruler B. , Verčič D. , Sriramesh K. . 2007 . ‘Defining strategic communication ,’ International Journal of Strategic Communication 1 : 3 – 35 . Google Scholar Crossref Search ADS Hart C. 2014 . Discourse, Grammar and Ideology: Functional and Cognitive Perspectives . Bloomsbury . Hooghiemstra R. 2000 . ‘Corporate communication and impression management–new perspectives why companies engage in corporate social reporting ,’ Journal of Business Ethics 27 : 55 – 68 . Google Scholar Crossref Search ADS Hunston S. , Thompson G. . 2000 . Evaluation in Text: Authorial Stance and the Construction of Discourse . Oxford University Press . Hyland K. 1996 . ‘Writing without conviction? Hedging in science research articles ,’ Applied Linguistics 17 : 433 – 54 . Google Scholar Crossref Search ADS Hyland K. 1998 . ‘Exploring corporate rhetoric: Metadiscourse in the CEO’s letter ,’ Journal of Business Communication 35 : 224 – 44 . Google Scholar Crossref Search ADS Jaworska S. , Nanda A. . 2016 . ‘Doing well by talking good: A topic modelling-assisted discourse study of corporate social responsibility ,’ Applied Linguistics 39 : 373 – 99 . Johansen T. S. , Ellerup Nielsen A. . 2011 . ‘Strategic stakeholder dialogues: A discursive perspective on relationship building ,’ Corporate Communications: An International Journal 16 : 204 – 17 . Google Scholar Crossref Search ADS Koller V. 2009 . ‘Corporate self-presentation and self-centredness: A case for cognitive critical discourse analysis’ in Pishwa H. (ed.): Language and Social Cognition: Expression of the Social Mind . De Gruyter , pp. 267 – 287 . KPMG . 2013 . International survey of corporate sustainability reporting. Available at http://www.kpmg.com/. Accessed 20 May 2016. Laine M. 2005 . ‘Meanings of the term ‘sustainable development’ in Finnish corporate disclosures ,’ Accounting Forum 29 : 395 – 413 . Google Scholar Crossref Search ADS Lischinsky A. 2011 . ‘In times of crisis: A corpus approach to the construction of the global financial crisis in annual reports ,’ Critical Discourse Studies 8 : 153 – 68 . Google Scholar Crossref Search ADS Livesey S. , Kearins K. . 2002 . ‘Transparent and caring corporations? A study of sustainability reports by the body shop and royal Dutch/Shell ,’ Organization and Environment 15 : 233 – 58 . Google Scholar Crossref Search ADS Malavasi D. 2010 . ‘The multifaceted nature of banks annual reports as informative, promotional and corporate communication practices’ in Allori P. Evangelisti , Garzone G. (eds): Discourse, Identities and Genres in Corporate Communication . Peter Lang , pp. 211 – 33 . Malavasi D. 2011 . ‘Doing well by doing good: A comparative analysis of Nokia’s and Ericcson’s corporate social responsibility reports’ in Grazone G. , Gotti M. (eds): Discourse, Communication and the Enterprise: Genres and Trends . Peter Lang , pp. 193 – 212 . Martin J. , White P. . 2005 . The Language of Evaluation: Appraisal in English . Palgrave Macmillan . McWilliams A. , Siegel D. S. , Wright P. M. . 2006 . ‘Corporate social responsibility: Strategic implications ,’ Journal of Management Studies 43 : 1 – 18 . Google Scholar Crossref Search ADS Merkl-Davies D. M. , Brennan N. M. . 2007 . ‘Discretionary disclosure strategies in corporate narratives: Incremental information or impression management? ,’ Journal of Accounting Literature 26 : 116 – 96 . Milne M. , Kearins K. , Walton S. . 2005 . ‘Playing with magic lanterns: The New Zealand Business Council for sustainable development and corporate triple bottom line reporting’ in Proc eedings of 4 APIRA Conference , July 4-6. Available at https://ourarchive.otago.ac.nz/handle/10523/1556. Accessed 20 May 2016. Milne M. , Tregidga H. , Walton S. . 2009 . ‘Words not actions! The ideological role of sustainable development reporting ,’ Accounting, Auditing and Accountability Journal 22 : 1211 – 57 . Google Scholar Crossref Search ADS Neu D. , Warsame H. , Pedwell K. . 1998 . ‘Managing public impressions: Environmental disclosures in annual reports ,’ Accounting, Organizations and Society 23 : 265 – 82 . Google Scholar Crossref Search ADS O’Donovan G. 2002 . ‘Environmental disclosures in the annual report: Extending the applicability and predictive power of legitimacy theory ,’ Accounting, Auditing and Accountability Journal 15 : 344 – 71 . Google Scholar Crossref Search ADS Ochs E. , Schiefelin B. . 1989 . ‘Language has a heart ,’ Text and Talk 9 : 7 – 26 . Palmer F. 2001 . Mood and Modality . Cambridge University Press . Pirson M. , Malhotra D. . 2011 . ‘Foundations of organizational trust: What matters to different stakeholders? ,’ Organization Science 22 : 1087 – 104 . Google Scholar Crossref Search ADS Rayson P. , Garside R. . 2000 . ‘Comparing corpora using frequency profiling’ in Proceedings of the Workshop on Comparing Corpora, held in Conjunction with the 38th Annual Meeting of the Association for Computational Linguistics , 1-8 October 2000, pp. 1–6. Available at http://ucrel.lancs.ac.uk/llwizard.html. Accessed 20 May 2016. Rutherford B. A. 2005 . ‘Genre analysis of corporate annual report narratives a corpus linguistics-based approach ,’ Journal of Business Communication 42 : 349 – 78 . Google Scholar Crossref Search ADS Skarmeas D. , Leonidou C. N. . 2013 . ‘When consumers doubt, watch out! The role of CSR skepticism ,’ Journal of Business Research 66 : 1831 – 8 . Google Scholar Crossref Search ADS Skulstad A. 2002 . Established and Emerging Business Genres . Norwegian Academic Press . Skulstad A. 2008 . ‘Creating a “green” image in the public sphere: Corporate environmental reports in a genre perspective’ in Wodak R. , Koller V. (eds): Handbook of Communication in the Public Sphere . De Gruyter , pp. 181 – 201 . Thompson G. , Hunston S. . 2000 . ‘Evaluation: An introduction’ in Hunston S. , Thompson G. (eds): Evaluation in Text: Authorial Stance and the Construction of Discourse . Oxford University Press , pp. 1 – 27 . Turker D. 2009 . ‘Measuring corporate social responsibility: A scale development study ,’ Journal of Business Ethics 85 : 411 – 27 . Google Scholar Crossref Search ADS Van Riel C. 1995 . Principles of Corporate Communication . Prentice Hall . Walker K. 2010 . ‘A systematic review of the corporate reputation literature: Definition, measurement, and theory ,’ Corporate Reputation Review 12 : 357 – 87 . Google Scholar Crossref Search ADS Westcott Alessandri S. 2001 . ‘Modeling corporate identity: A concept explication and theoretical explanation ,’ Corporate Communications: An International Journal 6 : 173 – 82 . Google Scholar Crossref Search ADS Appendix Table A1: List of stance markers used for automatic retrieval (step 1 of the analysis) Stance type Construction type Markers Attitudinal stance Ability or willingness adjective + post-predicate complement clause able, bound, committed, dedicated, determined, eager, unable, willing Attitude adverb alarmingly, amazingly, amusingly, annoyingly, ashamedly, astonishingly, conveniently, curiously, depressingly, disappointingly, disgustingly, disturbingly, embarrassedly, enchantingly, fortunately, frighteningly, funnily, happily, hopefully, impatiently, incredibly, inevitably, interestingly, ironically, luckily, mercifully, naturally, oddly, perplexingly, predictably, preferably, proudly, refreshingly, regretfully, regrettably, remarkably, rightly, sadly, shockingly, strangely, surprisingly, suspiciously, thankfully, tragically, unaccountably, understandably, unexpectedly, unfortunately, unhappily, unluckily, unnaturally Attitude verb + post-predicate complement clausea aggravates, agitates, alarms, amazes, amuses, annoys, astonishes, begrudge, bothers, confuses, deign, delights, despise, detest, disappoints, discourages, disgusts, dislike, dismays, distresses, disturbs, dread, embarrasses, enjoy, envy, fancy, fear, frightens, hate, hope, horrifies, interests, irritates, kills, like, loathe, love, overwhelms, pains, perplexes, perturbs, pleases, prefer, puzzles, refreshes, regret, relish, resent, rubs, saddens, scares, scorns, shocks, slays, surprises, thrills, troubles, upsets, worry Desire/intention/decision verb + to-complement clause aim, commit, crave, dedicate, intend, look, mean, necessitate, need, plan, require, seek, strive, want, wish, would like Ease or difficulty adjective + to-complement clause arduous, complex, complicated, demanding, difficult, easy, hard, impossible, impracticable, (not) possible, simple, straightforward, uncomplicated, undemanding Emotion/attitude adjective + post-predicate complement clause afraid, alarmed, amazed, ashamed, concerned, delighted, depressed, disappointed, disgusted, dismayed, dissatisfied, distressed, disturbed, eager, embarrassed, enchanted, encouraged, fascinated, frightened, furious, glad, happy, hopeful, impatient, indignant, interested, irritated, jubilant, keen, mad, merciful, overjoyed, overwhelmed, perplexed, perturbed, pleased, proud, puzzled, regretful, relieved, sad, scared, shocked, surprised, thankful, unexpected, unfortunate, unhappy, worried Evaluation adjective + post-predicate complement clause alarming, amazing, amusing, annoying, appropriate, astonishing, bad, confusing, convenient, critical, crucial, delightful, disappointing, disgusting, distressing, disturbing, embarrassing, encouraging, essential, fascinating, fortunate, frightening, fundamental, funny, good, gratifying, helpful, horrible, imperative, important, improper, inappropriate, inconceivable, incredible, inevitable, interesting, interesting, ironic, irritating, key, lucky, necessary, nice, notable, noteworthy, odd, paramount, perplexing, pleasing, positive, predictable, preferable, puzzling, reasonable, regrettable, ridiculous, scary, significant, silly, strange, surprising, suspicious, terrible, tragic, unacceptable, understandable, unexpected, unfortunate, unnatural, unsetting, vital, worrisome, worrying Epistemic stance Certainty adjective + post-predicate complement clause apparent, certain, clear, confident, convinced, definite, evident, impossible, inconceivable, incontestable, incontrovertible, indisputable, indubitable, irrefutable, manifest, not possible, obvious, patent, plain, positive, sure, true, unambiguous, unarguable, undeniable, undoubted, unmistakable, unquestionable, untrue, well-known Likelihood adjective + post-predicate complement clause alleged, arguable, conceivable, disputable, doubtful, dubious, imaginable, improbable, indefinite, likely, not certain, not clear, not sure, possible, presumable, probable, questionable, reputed, seeming, supposed, uncertain, unclear, unlikely, unsure Certainty adverb actually, admittedly, assuredly, avowedly, certainly, (in/with) certainty, clearly, (of) course, decidedly, definitely, (without) doubt, doubtlessly, evidently, (in) fact, incontrovertibly, indeed, indisputably, irrefutably, manifestly, obviously, patently, plainly, (in) reality, surely, unambiguously, unarguably, undeniably, unequivocally, unquestionably, veritably Likelihood adverb allegedly, apparently, arguably, conceivably, formally, hypothetically, ideally, likely, maybe, officially, ostensibly, outwardly, perchance, perhaps, possibly, potentially, presumably, purportedly, reportedly, reputedly, seemingly, superficially, supposedly, technically, theoretically, unlikely Certainty noun + that-complement clause certainty, conclusion, conviction, deduction, fact, finding, indication, judgment, knowledge, no doubt, observation, prediction, proof, resolution, sign Likelihood noun + that-complement clause assumption, belief, claim, doubt, estimate, feeling, guess, hypothesis, inference, probability, speculation, suspect, thought, view Certainty verb + post-predicate complement clause anticipate, ascertain, calculate, conclude, confirm, corroborate, deduce, demonstrate, determine, discern, establish, find, know, note, perceive, project, prove, ratified, realize, show, signify, understand, verify Likelihood verb + post-predicate complement clause appear, assume, believe, conjecture, consider, disbelieve, doubt, estimate, expect, feel, gather, guess, hypothesize, hypothesize, imagine, indicate, infer, postulate, presume, reckon, seem, sense, speculate, suggest, suppose, suspect, think Modality Permission/possibility/ability modal can, could, may, might Necessity/obligation modal must, should Volition/prediction modal shall, will, would Stance type Construction type Markers Attitudinal stance Ability or willingness adjective + post-predicate complement clause able, bound, committed, dedicated, determined, eager, unable, willing Attitude adverb alarmingly, amazingly, amusingly, annoyingly, ashamedly, astonishingly, conveniently, curiously, depressingly, disappointingly, disgustingly, disturbingly, embarrassedly, enchantingly, fortunately, frighteningly, funnily, happily, hopefully, impatiently, incredibly, inevitably, interestingly, ironically, luckily, mercifully, naturally, oddly, perplexingly, predictably, preferably, proudly, refreshingly, regretfully, regrettably, remarkably, rightly, sadly, shockingly, strangely, surprisingly, suspiciously, thankfully, tragically, unaccountably, understandably, unexpectedly, unfortunately, unhappily, unluckily, unnaturally Attitude verb + post-predicate complement clausea aggravates, agitates, alarms, amazes, amuses, annoys, astonishes, begrudge, bothers, confuses, deign, delights, despise, detest, disappoints, discourages, disgusts, dislike, dismays, distresses, disturbs, dread, embarrasses, enjoy, envy, fancy, fear, frightens, hate, hope, horrifies, interests, irritates, kills, like, loathe, love, overwhelms, pains, perplexes, perturbs, pleases, prefer, puzzles, refreshes, regret, relish, resent, rubs, saddens, scares, scorns, shocks, slays, surprises, thrills, troubles, upsets, worry Desire/intention/decision verb + to-complement clause aim, commit, crave, dedicate, intend, look, mean, necessitate, need, plan, require, seek, strive, want, wish, would like Ease or difficulty adjective + to-complement clause arduous, complex, complicated, demanding, difficult, easy, hard, impossible, impracticable, (not) possible, simple, straightforward, uncomplicated, undemanding Emotion/attitude adjective + post-predicate complement clause afraid, alarmed, amazed, ashamed, concerned, delighted, depressed, disappointed, disgusted, dismayed, dissatisfied, distressed, disturbed, eager, embarrassed, enchanted, encouraged, fascinated, frightened, furious, glad, happy, hopeful, impatient, indignant, interested, irritated, jubilant, keen, mad, merciful, overjoyed, overwhelmed, perplexed, perturbed, pleased, proud, puzzled, regretful, relieved, sad, scared, shocked, surprised, thankful, unexpected, unfortunate, unhappy, worried Evaluation adjective + post-predicate complement clause alarming, amazing, amusing, annoying, appropriate, astonishing, bad, confusing, convenient, critical, crucial, delightful, disappointing, disgusting, distressing, disturbing, embarrassing, encouraging, essential, fascinating, fortunate, frightening, fundamental, funny, good, gratifying, helpful, horrible, imperative, important, improper, inappropriate, inconceivable, incredible, inevitable, interesting, interesting, ironic, irritating, key, lucky, necessary, nice, notable, noteworthy, odd, paramount, perplexing, pleasing, positive, predictable, preferable, puzzling, reasonable, regrettable, ridiculous, scary, significant, silly, strange, surprising, suspicious, terrible, tragic, unacceptable, understandable, unexpected, unfortunate, unnatural, unsetting, vital, worrisome, worrying Epistemic stance Certainty adjective + post-predicate complement clause apparent, certain, clear, confident, convinced, definite, evident, impossible, inconceivable, incontestable, incontrovertible, indisputable, indubitable, irrefutable, manifest, not possible, obvious, patent, plain, positive, sure, true, unambiguous, unarguable, undeniable, undoubted, unmistakable, unquestionable, untrue, well-known Likelihood adjective + post-predicate complement clause alleged, arguable, conceivable, disputable, doubtful, dubious, imaginable, improbable, indefinite, likely, not certain, not clear, not sure, possible, presumable, probable, questionable, reputed, seeming, supposed, uncertain, unclear, unlikely, unsure Certainty adverb actually, admittedly, assuredly, avowedly, certainly, (in/with) certainty, clearly, (of) course, decidedly, definitely, (without) doubt, doubtlessly, evidently, (in) fact, incontrovertibly, indeed, indisputably, irrefutably, manifestly, obviously, patently, plainly, (in) reality, surely, unambiguously, unarguably, undeniably, unequivocally, unquestionably, veritably Likelihood adverb allegedly, apparently, arguably, conceivably, formally, hypothetically, ideally, likely, maybe, officially, ostensibly, outwardly, perchance, perhaps, possibly, potentially, presumably, purportedly, reportedly, reputedly, seemingly, superficially, supposedly, technically, theoretically, unlikely Certainty noun + that-complement clause certainty, conclusion, conviction, deduction, fact, finding, indication, judgment, knowledge, no doubt, observation, prediction, proof, resolution, sign Likelihood noun + that-complement clause assumption, belief, claim, doubt, estimate, feeling, guess, hypothesis, inference, probability, speculation, suspect, thought, view Certainty verb + post-predicate complement clause anticipate, ascertain, calculate, conclude, confirm, corroborate, deduce, demonstrate, determine, discern, establish, find, know, note, perceive, project, prove, ratified, realize, show, signify, understand, verify Likelihood verb + post-predicate complement clause appear, assume, believe, conjecture, consider, disbelieve, doubt, estimate, expect, feel, gather, guess, hypothesize, hypothesize, imagine, indicate, infer, postulate, presume, reckon, seem, sense, speculate, suggest, suppose, suspect, think Modality Permission/possibility/ability modal can, could, may, might Necessity/obligation modal must, should Volition/prediction modal shall, will, would Note: aThe verbs reported in the third person were considered as stance markers only when appearing in the construction ‘it (ADV) ____ that/to’. Conversely, for the verbs reported in the base form, all inflected forms were considered. Table A1: List of stance markers used for automatic retrieval (step 1 of the analysis) Stance type Construction type Markers Attitudinal stance Ability or willingness adjective + post-predicate complement clause able, bound, committed, dedicated, determined, eager, unable, willing Attitude adverb alarmingly, amazingly, amusingly, annoyingly, ashamedly, astonishingly, conveniently, curiously, depressingly, disappointingly, disgustingly, disturbingly, embarrassedly, enchantingly, fortunately, frighteningly, funnily, happily, hopefully, impatiently, incredibly, inevitably, interestingly, ironically, luckily, mercifully, naturally, oddly, perplexingly, predictably, preferably, proudly, refreshingly, regretfully, regrettably, remarkably, rightly, sadly, shockingly, strangely, surprisingly, suspiciously, thankfully, tragically, unaccountably, understandably, unexpectedly, unfortunately, unhappily, unluckily, unnaturally Attitude verb + post-predicate complement clausea aggravates, agitates, alarms, amazes, amuses, annoys, astonishes, begrudge, bothers, confuses, deign, delights, despise, detest, disappoints, discourages, disgusts, dislike, dismays, distresses, disturbs, dread, embarrasses, enjoy, envy, fancy, fear, frightens, hate, hope, horrifies, interests, irritates, kills, like, loathe, love, overwhelms, pains, perplexes, perturbs, pleases, prefer, puzzles, refreshes, regret, relish, resent, rubs, saddens, scares, scorns, shocks, slays, surprises, thrills, troubles, upsets, worry Desire/intention/decision verb + to-complement clause aim, commit, crave, dedicate, intend, look, mean, necessitate, need, plan, require, seek, strive, want, wish, would like Ease or difficulty adjective + to-complement clause arduous, complex, complicated, demanding, difficult, easy, hard, impossible, impracticable, (not) possible, simple, straightforward, uncomplicated, undemanding Emotion/attitude adjective + post-predicate complement clause afraid, alarmed, amazed, ashamed, concerned, delighted, depressed, disappointed, disgusted, dismayed, dissatisfied, distressed, disturbed, eager, embarrassed, enchanted, encouraged, fascinated, frightened, furious, glad, happy, hopeful, impatient, indignant, interested, irritated, jubilant, keen, mad, merciful, overjoyed, overwhelmed, perplexed, perturbed, pleased, proud, puzzled, regretful, relieved, sad, scared, shocked, surprised, thankful, unexpected, unfortunate, unhappy, worried Evaluation adjective + post-predicate complement clause alarming, amazing, amusing, annoying, appropriate, astonishing, bad, confusing, convenient, critical, crucial, delightful, disappointing, disgusting, distressing, disturbing, embarrassing, encouraging, essential, fascinating, fortunate, frightening, fundamental, funny, good, gratifying, helpful, horrible, imperative, important, improper, inappropriate, inconceivable, incredible, inevitable, interesting, interesting, ironic, irritating, key, lucky, necessary, nice, notable, noteworthy, odd, paramount, perplexing, pleasing, positive, predictable, preferable, puzzling, reasonable, regrettable, ridiculous, scary, significant, silly, strange, surprising, suspicious, terrible, tragic, unacceptable, understandable, unexpected, unfortunate, unnatural, unsetting, vital, worrisome, worrying Epistemic stance Certainty adjective + post-predicate complement clause apparent, certain, clear, confident, convinced, definite, evident, impossible, inconceivable, incontestable, incontrovertible, indisputable, indubitable, irrefutable, manifest, not possible, obvious, patent, plain, positive, sure, true, unambiguous, unarguable, undeniable, undoubted, unmistakable, unquestionable, untrue, well-known Likelihood adjective + post-predicate complement clause alleged, arguable, conceivable, disputable, doubtful, dubious, imaginable, improbable, indefinite, likely, not certain, not clear, not sure, possible, presumable, probable, questionable, reputed, seeming, supposed, uncertain, unclear, unlikely, unsure Certainty adverb actually, admittedly, assuredly, avowedly, certainly, (in/with) certainty, clearly, (of) course, decidedly, definitely, (without) doubt, doubtlessly, evidently, (in) fact, incontrovertibly, indeed, indisputably, irrefutably, manifestly, obviously, patently, plainly, (in) reality, surely, unambiguously, unarguably, undeniably, unequivocally, unquestionably, veritably Likelihood adverb allegedly, apparently, arguably, conceivably, formally, hypothetically, ideally, likely, maybe, officially, ostensibly, outwardly, perchance, perhaps, possibly, potentially, presumably, purportedly, reportedly, reputedly, seemingly, superficially, supposedly, technically, theoretically, unlikely Certainty noun + that-complement clause certainty, conclusion, conviction, deduction, fact, finding, indication, judgment, knowledge, no doubt, observation, prediction, proof, resolution, sign Likelihood noun + that-complement clause assumption, belief, claim, doubt, estimate, feeling, guess, hypothesis, inference, probability, speculation, suspect, thought, view Certainty verb + post-predicate complement clause anticipate, ascertain, calculate, conclude, confirm, corroborate, deduce, demonstrate, determine, discern, establish, find, know, note, perceive, project, prove, ratified, realize, show, signify, understand, verify Likelihood verb + post-predicate complement clause appear, assume, believe, conjecture, consider, disbelieve, doubt, estimate, expect, feel, gather, guess, hypothesize, hypothesize, imagine, indicate, infer, postulate, presume, reckon, seem, sense, speculate, suggest, suppose, suspect, think Modality Permission/possibility/ability modal can, could, may, might Necessity/obligation modal must, should Volition/prediction modal shall, will, would Stance type Construction type Markers Attitudinal stance Ability or willingness adjective + post-predicate complement clause able, bound, committed, dedicated, determined, eager, unable, willing Attitude adverb alarmingly, amazingly, amusingly, annoyingly, ashamedly, astonishingly, conveniently, curiously, depressingly, disappointingly, disgustingly, disturbingly, embarrassedly, enchantingly, fortunately, frighteningly, funnily, happily, hopefully, impatiently, incredibly, inevitably, interestingly, ironically, luckily, mercifully, naturally, oddly, perplexingly, predictably, preferably, proudly, refreshingly, regretfully, regrettably, remarkably, rightly, sadly, shockingly, strangely, surprisingly, suspiciously, thankfully, tragically, unaccountably, understandably, unexpectedly, unfortunately, unhappily, unluckily, unnaturally Attitude verb + post-predicate complement clausea aggravates, agitates, alarms, amazes, amuses, annoys, astonishes, begrudge, bothers, confuses, deign, delights, despise, detest, disappoints, discourages, disgusts, dislike, dismays, distresses, disturbs, dread, embarrasses, enjoy, envy, fancy, fear, frightens, hate, hope, horrifies, interests, irritates, kills, like, loathe, love, overwhelms, pains, perplexes, perturbs, pleases, prefer, puzzles, refreshes, regret, relish, resent, rubs, saddens, scares, scorns, shocks, slays, surprises, thrills, troubles, upsets, worry Desire/intention/decision verb + to-complement clause aim, commit, crave, dedicate, intend, look, mean, necessitate, need, plan, require, seek, strive, want, wish, would like Ease or difficulty adjective + to-complement clause arduous, complex, complicated, demanding, difficult, easy, hard, impossible, impracticable, (not) possible, simple, straightforward, uncomplicated, undemanding Emotion/attitude adjective + post-predicate complement clause afraid, alarmed, amazed, ashamed, concerned, delighted, depressed, disappointed, disgusted, dismayed, dissatisfied, distressed, disturbed, eager, embarrassed, enchanted, encouraged, fascinated, frightened, furious, glad, happy, hopeful, impatient, indignant, interested, irritated, jubilant, keen, mad, merciful, overjoyed, overwhelmed, perplexed, perturbed, pleased, proud, puzzled, regretful, relieved, sad, scared, shocked, surprised, thankful, unexpected, unfortunate, unhappy, worried Evaluation adjective + post-predicate complement clause alarming, amazing, amusing, annoying, appropriate, astonishing, bad, confusing, convenient, critical, crucial, delightful, disappointing, disgusting, distressing, disturbing, embarrassing, encouraging, essential, fascinating, fortunate, frightening, fundamental, funny, good, gratifying, helpful, horrible, imperative, important, improper, inappropriate, inconceivable, incredible, inevitable, interesting, interesting, ironic, irritating, key, lucky, necessary, nice, notable, noteworthy, odd, paramount, perplexing, pleasing, positive, predictable, preferable, puzzling, reasonable, regrettable, ridiculous, scary, significant, silly, strange, surprising, suspicious, terrible, tragic, unacceptable, understandable, unexpected, unfortunate, unnatural, unsetting, vital, worrisome, worrying Epistemic stance Certainty adjective + post-predicate complement clause apparent, certain, clear, confident, convinced, definite, evident, impossible, inconceivable, incontestable, incontrovertible, indisputable, indubitable, irrefutable, manifest, not possible, obvious, patent, plain, positive, sure, true, unambiguous, unarguable, undeniable, undoubted, unmistakable, unquestionable, untrue, well-known Likelihood adjective + post-predicate complement clause alleged, arguable, conceivable, disputable, doubtful, dubious, imaginable, improbable, indefinite, likely, not certain, not clear, not sure, possible, presumable, probable, questionable, reputed, seeming, supposed, uncertain, unclear, unlikely, unsure Certainty adverb actually, admittedly, assuredly, avowedly, certainly, (in/with) certainty, clearly, (of) course, decidedly, definitely, (without) doubt, doubtlessly, evidently, (in) fact, incontrovertibly, indeed, indisputably, irrefutably, manifestly, obviously, patently, plainly, (in) reality, surely, unambiguously, unarguably, undeniably, unequivocally, unquestionably, veritably Likelihood adverb allegedly, apparently, arguably, conceivably, formally, hypothetically, ideally, likely, maybe, officially, ostensibly, outwardly, perchance, perhaps, possibly, potentially, presumably, purportedly, reportedly, reputedly, seemingly, superficially, supposedly, technically, theoretically, unlikely Certainty noun + that-complement clause certainty, conclusion, conviction, deduction, fact, finding, indication, judgment, knowledge, no doubt, observation, prediction, proof, resolution, sign Likelihood noun + that-complement clause assumption, belief, claim, doubt, estimate, feeling, guess, hypothesis, inference, probability, speculation, suspect, thought, view Certainty verb + post-predicate complement clause anticipate, ascertain, calculate, conclude, confirm, corroborate, deduce, demonstrate, determine, discern, establish, find, know, note, perceive, project, prove, ratified, realize, show, signify, understand, verify Likelihood verb + post-predicate complement clause appear, assume, believe, conjecture, consider, disbelieve, doubt, estimate, expect, feel, gather, guess, hypothesize, hypothesize, imagine, indicate, infer, postulate, presume, reckon, seem, sense, speculate, suggest, suppose, suspect, think Modality Permission/possibility/ability modal can, could, may, might Necessity/obligation modal must, should Volition/prediction modal shall, will, would Note: aThe verbs reported in the third person were considered as stance markers only when appearing in the construction ‘it (ADV) ____ that/to’. Conversely, for the verbs reported in the base form, all inflected forms were considered. © Oxford University Press 2017 This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

Applied LinguisticsOxford University Press

Published: Dec 1, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off