Introduction

Corporate social responsibilities are increasingly extended beyond the boundaries of the firm (Gold et al. 2010). Particularly non-governmental organizations often make claims suggesting that firms are responsible for harm that occurs anywhere in the supply chain. Firms develop non-market strategies to respond to such claims from civil society players (Baron 1995). Traditional approaches to corporate social responsibility (CSR) have connected social responsibilities with those parties who have a legitimate stake in the company (e.g., Freeman 2004, 1984).

However, such stakeholder approaches do not always fully capture responsibilities along a global supply chain. Some approaches have responded by extending the stakeholder conception in a way that NGOs may qualify as stakeholders (e.g., Doh and Teegen 2002; Teegen et al. 2004; Baur and Palazzo 2011). Others have emphasized the dependence of stakeholder approaches on the nation state as an analytical reference (Scherer and Palazzo 2011) and suggested a more political role for the firm (Scherer and Palazzo 2011; Baur and Palazzo 2011; Baur and Schmitz 2012).

We propose that a cosmopolitan approach to supply chain responsibility (SCR) activity may enhance our understanding by suggesting that claims are legitimized by citizenship rather than by having a stake in a company. Following Hayden (2005), we distinguish between moralistic and legalistic cosmopolitanism. While the latter seeks to protect people’s rights via the development of formalized rules and may thus limit responsibilities for others to the extent that is specified in those rules, the former suggests that responsibilities result from the moral relationships we have with others by virtue of our common humanity (Hayden 2005, p. 3). Both moralistic and legalistic cosmopolitanism can thus help in understanding claims for firms’ responsibilities for their potentially global, loose, and unstructured supply chains.

In the paper below, we aim to contribute to the understanding of firms’ and NGOs’ SCR discourses. We ask if different forms of cosmopolitan language, moralistic and legalistic, can be identified in the language used by firms and NGOs in their SCR discourse. We further aim to understand if the use of these different forms of cosmopolitan language connects to firm realities and if it can be connected to organizations’ understanding of firms’ role in the SCR debate.

Our paper continues in section two by reviewing the difficulties in applying stakeholder approaches to supply chain responsibility claims, and then introducing cosmopolitanism as an approach that takes CSR beyond the boundaries of the firm to demand new forms of firm involvement. The development of new norms that attribute responsibility on the basis of social connectedness rather than direct liability (Young 2008) transforms supplier relationships, making them the object of broader, social concern. We distinguish a legalistic from a moralistic approach to cosmopolitanism and connect them with SCR. In section three, we introduce our material, comprising NGO and firm SCR reports that address issues of supply chain responsibility, with a focus on claims raised by NGOs regarding water pollution in China. We then introduce our methodology, which brings together qualitative and quantitative linguistic techniques supported by interview data. In section four, we present our findings, noting two patterns of cosmopolitan discourse, a legalistic and a moralistic one, that appear in our data. Rather than opposed types, we perceive these forms of discourse to exist along a continuum. We discuss our findings in section five. Before concluding the paper, section six presents managerial implications for firms’ non-market strategies and discusses limitations of our work together with avenues for further research.

Theoretical Framework: Cosmopolitanism and Supply Chain Responsibility

The Problem of Defining Responsibilities in the Supply Chain

In recent years, there has been an expansion of claims on firms to accept responsibility for activities in their supply chains. While NGOs have traditionally addressed the activities of firms themselves (e.g., NGO criticism of Nestlé’s marketing of breast milk substitutes or of Shell’s decision to sink the Brent Spar oil storage buoy), the expansion of outsourcing activities and the transformation of many firms from producers to brand managers has resulted in an increasing tendency for NGOs to attribute responsibility to firms for practices and behaviors that are actually undertaken by others, such as suppliers. These claims seem to be about the focal firm’s potential influence rather than its own actions, and typically involve demands for the firm to make its supplier relations transparent (Doorey 2011) and take responsibility for activities anywhere in its supply chain (Anderson and Skjoett-Larsen 2009).

This suggests a type of cosmopolitan thinking that attributes responsibility not on the basis of a direct stakeholder relationship, but instead on the basis of a common humanity. This conception of responsibility, which is based on a ‘social connection model,’ can be contrasted with the ‘liability model’ that assigns responsibility to those who can be directly linked to the harms they cause (Young 2008). Showing how anti-sweatshop activists have taken a social connection approach, Young (2008) argues that in this model an agent’s structural position and associated level of power or influence is a key factor that determines its degree of responsibility for harm. In the case of supply chains, power is typically held by brands or retailers, who in this model therefore bear responsibility, even though they are not directly legally responsible for the actions of suppliers.

Claims regarding SCR raise the question of what the boundaries and limits of these responsibilities should be and how they can be implemented (Mares 2010; Scherer and Palazzo 2011; Amaeshi et al. 2008). Attempts to specify and define these limits have become a focus of NGO-firm discursive engagement in recent years. This discussion is also connected to the deeper-rooted question of the role of the firm and whether firms’ role remains restricted to value maximization or goes beyond that. In this context, Crane et al. (2008) suggested that firms may take roles as citizens or governments, and these roles connect to the responsibilities one would attribute to firms for their supply chains.

One way in which this discussion has continued is through the publication of various NGO and company reports on supply chain issues. Below we first depict how stakeholder approaches have been used to address NGO-firm engagement. However, we argue that applying stakeholder theory to supply chain responsibility is inappropriate and contradicts its character as an ethical theory of organizational management. We then propose cosmopolitan approaches as an alternative and argue that text analyses of reports are a suitable method to explore the issue further.

The Limitations of Stakeholder Approaches: Responsibility Based on Relationships

Existing research often approaches firm-NGO engagements from the perspective of stakeholder theory (Yaziji and Doh 2009), and it might therefore appear logical to examine SCR discourse from this viewpoint. As authors who have developed stakeholder theory are careful to point out, this approach is a theory of organizational ethics and not a comprehensive ethical theory (Phillips et al. 2003). While stakeholder theories justify obligations that firms (and people working within them) have toward stakeholders, it neglects other, more fundamental moral obligations that firms have to all people, regardless of stakeholder status. These latter obligations are rooted in universal ‘hypernorms’ of ethical responsibility (Donaldson and Dunfee 1994). Stakeholder approaches thus fall short in providing guidance on how to balance the needs of stakeholders with those who do not have a stake in the firm (Phillips et al. 2003). We thus argue that the utility of a stakeholder perspective for SCR discourse analysis remains limited for two reasons. First, NGOs’ status as stakeholders seems problematic, and second, the SCR discourse may go beyond claims regarding specific company practices and instead attempt to define new roles and responsibilities for the firm. We elaborate on these two aspects.

First, when an NGO criticizes a firm for environmental or labor rights violations allegedly committed by its suppliers, it seems difficult to argue that the NGO actually has a stake in the focal firm. However, the concept of “derivative stakeholders” has been suggested to address this concern (Phillips 2003). Our second argument seems more significant: supply chain responsibility discourse tends to go beyond specific firm policies to instead consider new roles for the firm and an expansion of its responsibilities. When NGOs make claims about supply chain responsibility, they may call for a firm to transcend its normal, value-maximizing business approach to engage in other behaviors, such as rule development and enforcement (government-like behavior) or collaboration with civil society actors (citizen-like behavior) (Crane et al. 2008). Because such claims call into question the fundamental organizational objectives of the firm, they are difficult to conceptualize within a stakeholder framework. While stakeholder approaches focus on how firm decisions affect stakeholders, cosmopolitan approaches appear instead to address firms’ social role. We thus believe that cosmopolitan approaches are more suitable for analyzing SCR discourse.

Cosmopolitanism: Responsibility Based on Global Citizenship

Cosmopolitanism bases citizenship and thus responsible participation in a global community that is not limited by political borders but instead encompasses the entire world (Holton 2009; Held 2005; Archibugi 2008). While goals of building a single global community are not new, recent economic, social, and political developments point to an increasingly cosmopolitan reality (Beck 2006), and there is growing agreement that citizenship has become disassociated from nationality, and sovereignty is no longer defined entirely with reference to states (Delanty 2000, p. 53). Business activity has contributed to this (Crane et al. 2008): On the one hand, firms are considered as corporate citizens (e.g., Matten and Crane 2005; Waddock 2008); on the other hand, through firms’ engagement with NGOs a global public domain is being co-created in which new issues are being raised and discussed (Ruggie 2003). Firms increasingly speak about their activities and responsibilities in cosmopolitan terms (Garsten 2003).

There are various conceptions of cosmopolitanism (Holton 2009), but two main trends can be discerned: Legalistic cosmopolitanism is concerned with the construction of institutional measures to ensure the protection of the human rights of all persons. By contrast, moralistic cosmopolitanism is based on positive notions of the worth of each human. This position creates obligations to others and thereby limits people’s actions (Hayden 2005). While legalistic cosmopolitanism creates cosmopolitan responsibilities based on institutions and law, moralistic cosmopolitanism promotes conceptions of cosmopolitan responsibility based on a sense of common humanity.

In the SCR discourse, these two perspectives are also connected to different roles that firms are expected and willing to take with regard to their responsibilities as suggested by Crane et al. (2008). The moralistic cosmopolitan language supports the idea of the company in the role of a citizen with the responsibility to do good to the other members of the community. Legalistic cosmopolitan language use connects to firms’ taking of a governmental function. While both NGOs and firms using legalistic cosmopolitan discourse may subscribe to moralist claims, they generally describe firms’ role as limited to ensuring that rules and laws directed to support these aims are followed within their sphere of influence. E.g., if there are laws or voluntary codes that forbid the discharge of toxic chemicals, then it is the firms’ responsibility to assure that these laws and codes are followed in their supply chain. In this conception, firms may further contribute to the development of rules and laws.

In line with Basu and Palazzo (2008), we believe that language can indicate how organizations make sense of the world. Company and NGO discourses not only reflect the actors’ positions on an issue, but they simultaneously shape and construct the issue in question (Burchell and Cook 2006), and provide a means for actors to legitimate their actions (Joutsenvirta 2011; Joutsenvirta and Vaara 2008; Vaara and Tienari 2008). For instance, Joutsenvirta (2011) has used critical discourse analysis (CDA) to show how companies and NGOs jointly determine the parameters of appropriate corporate conduct. Similarly, company supply chain responsibility reporting—and NGO reports on supply chain problems—is a way in which companies and civil society actors establish their SCR discourses and shape firms’ responsibilities for their supply chains. In texts of this sort, a tension between legalistic and moralistic cosmopolitan approaches becomes evident. An example is provided by the following extract from a company supply chain responsibility report:

Workers everywhere should have the right to safe and ethical working conditions. They should also have access to educational opportunities to improve their lives. Through a continual cycle of inspections, improvement plans, and verification, we work with our suppliers to make sure they comply with our Code of Conduct and live up to these ideals. (Apple 2013, p. 3)

While the first and second sentences declare the firm’s moralist cosmopolitan aspirations, the third sentence limits the company’s role and responsibility substantially. Although the firm states its support of workers’ rights in universalizing terms, it defines its own responsibility in legalistic terms, as limited to a more governmental role of monitoring a key stakeholder—suppliers—to confirm that the latter realize these moralist cosmopolitan ‘ideals.’ Firms may thus combine moralist and legalist cosmopolitan language in their discourses. For better understanding firms’ positions, we apply a vocabulary structure approach (Loewenstein et al. 2012) to SCR reports of firms and NGOs in order to understand what the texts are mainly about and which cosmopolitan approach they thus mainly reflect.

Materials and Methods

In order to investigate the extent to which legalistic and moralistic cosmopolitan language can be identified in the SCR discourse of firms and NGOs, we analyze reports about supply chain responsibility in China produced by two MNCs and one international NGO and one group of Chinese NGOs associated with the Green Choice Alliance. As the “world’s factory,” China is a major site of outsourced production and thus a focus of supply chain activism. Moreover, with the second largest GDP in the world, China is also a critical market for many firms and therefore an important site for investigation. We first depict two industry-wide campaigns focusing specifically on the discourses of one MNC and one NGO or group of NGOs, and then explain our choice of these organizations and the related texts.

Apple and Chinese Environmental NGOs in the Electronics Sector

We study Apple, a global brand leader in the field of consumer electronics, and a coalition of Chinese environmental NGOs. In 2011, the NGOs published a series of two reports about Apple’s suppliers based on pollution information obtained under China’s environmental information disclosure law (Zhang and Barr 2013). This was part of an industry-wide campaign drawing attention to pollution in the IT sector. The NGOs claimed that Apple’s association with such suppliers indicated weaknesses in its supply chain auditing procedures and a failure to live up to its own supplier code of conduct. The Chinese NGOs further argued that Apple bore responsibility for these problems and was violating the promises made in its code (Friends of Nature (FON) et al. 2011a). After initial reluctance to engage with the NGOs, Apple held a series of face-to-face meetings with the groups which resulted in new environmental auditing procedures in China that involved the NGOs as partners (FON et al. 2013; Apple 2013). In 2013 the NGOs published Apple Opens Up, a report documenting the environmental improvements that had been achieved through firm-NGO cooperation, but also pushing Apple to take further action. Apple’s description of the relationship is presented in its Apple 2013 and 2014 Supplier Responsibility Progress Reports.

Adidas and Greenpeace in the Chinese Textile Sector

We further examine the industry-wide ‘detox’ campaign launched by Greenpeace International aimed at eliminating the use of toxic chemical inputs in the apparel supply chain. This campaign was supported with the Dirty Laundry report, which exposed toxic chemical use at two Chinese textile suppliers and criticized companies’ CSR policies for failing to prevent suppliers’ discharge of hazardous chemicals (Greenpeace 2011a). This report was followed by other reports including Dirty Laundry 2 (Greenpeace 2011b) and Dirty Laundry: Reloaded (Greenpeace 2012). Significantly, Greenpeace attributed responsibility to brands that used polluting suppliers, whether or not the brands were actually associated with any of the production processes causing chemical pollution. Several of the brands targeted by Greenpeace responded by establishing their own industry-led process for eliminating toxics, the “Roadmap to Zero” (Roadmap to Zero 2011).

Adidas was one of several companies targeted by Greenpeace’s campaign. Adidas responded to Greenpeace by pledging to eliminate toxic inputs (Dubsky 2011) and joining the “Roadmap to Zero” process. We explore the company’s response and the language used in its Sustainability Progress 2011, 2012, and 2013 reports.

Selection of Organizations and Materials

We aimed to locate two potentially contrasting forms of discourse at the firm level and at the NGO level. We therefore chose four organizations (two firms, an NGO, and a group of NGOs) that appeared to be distinct in their characteristics. We then chose to analyze their SCR reports with a particular interest in capturing the organizations’ general stance toward SCR, rather than campaign-specific responses. We explain this in detail.

Following the stakeholder culture continuum of Jones et al. (2007), we distinguished self-regarding from other-regarding firms, and identified Apple as a more self-regarding firm—this is supported by the analysis of Waldron et al. (2013). By contrast, we identify Adidas as a more ‘other-regarding’ firm, which is supported by research describing Adidas as a firm with a collaborative style of supplier partnership (Frenkel and Scott 2002) and a strong concern for supply chain pressures (Tate et al. 2010).

We consider the four 2011–2014 Supplier Responsibility Progress Reports for Apple and the three Sustainability Progress 2011–2013 reports for Adidas. We have selected these texts for analysis because we believe that they best demonstrate each organization’s position as well as its framing of its position. These reports are the texts that best indicate the ways in which the organizations make sense of SCR issues and present their position to a wider public. Moreover, it can be argued that these sorts of CSR communication practices do not simply represent the organization’s views, but substantially contribute to constituting the organization itself (Schoeneborn and Trittin 2013). For both firms the chosen documents thus represent at first general reports on supply chain responsibility and depict the firms’ general position regarding SCR. At the same time these reports indicate an awareness of the positions expressed by the NGOs and thus contain some elements of interactive communication. For example: “When Greenpeace launched its Detox Campaign … we worked together with a coalition of other brands to develop a joint roadmap…” (Adidas 2012, p. 4), and “… we work with independent organizations such as … the Institute of Public and Environmental Affairs (IPE).” (Apple 2013, p. 26). (Note: the IPE has played a major role in the Green Choice Alliance of Chinese environmental NGOs and is co-author of the reports published by FON et al.). In contrast to other recent studies, which have for example analyzed the dialogic character of communication between Greenpeace and companies mentioned in the Dirty Laundry report by examining press releases (e.g., Brennan et al. 2013), we chose company reports on SCR, because we consider them to be the most important documents for indicating an organization’s stance toward SCR and CSR, rather than a statement framed to respond to a very specific claim.

For the NGOs we referred to Yaziji and Doh’s (2009) distinction between watchdog NGOs and social movement NGOs. Watchdog NGOs tend not to pursue radical change and rather have a monitoring function, while social movement NGOs may aim at changing an aspect of the existing system. We consider a group of Chinese NGOs as watchdog NGOs. Greenpeace, in contrast bears the characteristics of a social movement NGO (den Hond 2010).

Methods

The corpus of documents analyzed includes a total of over 155,000 words (see Table 1). We used electronic text analysis (Adolphs 2006; Baker 2006; Hoey et al. 2007; Mautner 2009) to help identify discursive features. After ‘cleaning’ the texts to remove redundant material such as page headers, footnotes, and captions, we used the WordSmith Tools 5 software package to produce a variety of analyses that would allow us to understand the vocabulary structure of the texts. In particular, we performed word counts, keywords analyses, and concordances. These analyses provide information on the content of the reports as well as the discursive construction of concepts. While word counts indicate the raw numbers of words in a single document or set of documents, keywords analyses compare the frequency of a word in a target document or corpus with its representation in a reference corpus. This analysis produces a log likelihood score that suggests how over or underrepresented a word is with comparison to the reference corpus. Keywords analysis thus indicates what a text is ‘about’ (Mahlberg 2007). We used the Freiburg-Brown Corpus of American English (FROWN corpus) as a reference, which allowed us to compare the words in our texts with 1 million words of informative text in written American English (McEnery et al. 2006). Because some of the texts have been produced in an international context, we also ran our analyses using the FLOB corpus as a reference and found our below results broadly supported. The FLOB (Freiburg Lancaster Oslo Bergen corpus) is equivalent to the FROWN but uses British English texts. Concordances demonstrate how a focal word is used in context, and therefore allows for an exploration of word partners as well as semantic preferences, that is, the kinds of meanings that attach to categories of words. By revealing patterns of language use, concordances can help in identifying discourses (Baker 2006).

Table 1 Reports analyzed
Full size table

Our findings result from corpus-assisted analysis (Baker 2010), meaning that we did not approach the textual data with preformulated hypotheses. Rather we ran general keywords analyses that revealed patterns in both individual documents and groups of texts. We used these initial results to identify prominent words, which we then investigated more closely for patterns of use using concordance analysis. To improve our familiarity with the campaigns and better frame our text analysis (Fairclough 2003), we supplemented our electronic text analysis with further sources of data including reports from the Chinese and international press. Additionally, to confirm our understanding of the campaigns and the reports and to contextualize them, we conducted interviews with Greenpeace and representatives of Chinese NGOs involved in the campaigns. Further interviews were held with representatives of other civil society organizations that were knowledgeable about the campaigns, resulting in a total of 17 interviews (over 20 h of recorded material). Interviews were conducted in English, Chinese, and German, and were later transcribed and translated into English if necessary.

Findings

In our investigation of supply chain responsibility discourse, we find that organizations adopt distinct variants of cosmopolitan language. Specifically, following Hayden’s (2005) categorization, we see examples of legalistic and moralistic forms of cosmopolitan discourse. We find more examples of legalistic language in the reports of Apple and the Chinese NGOs and more moralistic language in the reports of Adidas and Greenpeace.

Chinese NGOs and Apple—Examples of More Legalistic Cosmopolitan Language

The data we analyzed includes the set of Apple’s Supplier Responsibility Progress Reports (SRPRs) from 2011 to 2014 as well as three reports about Apple jointly published by several Chinese NGOs between January 2011 and January 2013. Our three main results are summarized and elaborated below. First, keywords analysis indicates that the discourses of both the NGOs and the company is about standards, their violation and their enforcement (see Tables 2, 3). The focus is on identifying violations and resolving them. Second, concordance analysis suggests that Apple plays the role of rule enforcer, applying its code and requiring change at suppliers (see Table 4). Third, while the NGOs base their claims on local law as well as Apple’s code, Apple frequently refers to its own code (see Table 5).

Table 2 Keywords in Apple 2011–2014 SRPR
Full size table
Table 3 Keywords analysis of FON et al.’s reports on Apple, 2011a, 2011b, 2013
Full size table
Table 4 Examples of ‘we’ concordances from analysis of Apple 2011–2014 SRPR
Full size table
Table 5 Examples of ‘our’ concordances from analysis of Apple 2011–2014 SRPR
Full size table

First, an examination of the keywords of Apple’s 2011–2014 SRPRs suggests a strong legalistic focus. This is not surprising, as the function of the report is to provide audit results. ‘Audits’ (#8, i.e., 8th rank in the keyword list) and ‘audit’ (#9), a tool that is used to discover problems, are key words, as is ‘violations’ (#19). ‘Code’ also features as the 12th most prominent word. Other words related to rules and their enforcement include ‘underage’ (#16), as in the responsibility violation of underage workers. The Chinese NGOs’ reports highlight problems associated with suppliers’ factories such as ‘pollution’ (#4) and ‘discharge’ (#7) (Table 3). Solutions are also considered, and ‘audit’ is ranked 19th and thus less prominent than in Apple’s SRPRs.

A second striking aspect of Apple’s reports is its use of ‘we,’ which is extremely prominent in the text, ranking as the 5th keyword—followed by ‘our’, which is 6th (Table 2). Use of pronouns can reveal orientations toward status and affiliation (Fairclough 2010; Pennebaker 2011). ‘We’ is not typically used in informational texts, but rather in promotional texts (Bhatia 2004). Apple’s use of ‘we’ appears to be uniformly exclusive, meaning that it does not include the reader/addressee of the reports (Pennebaker 2011). Although the term may be used to personalize the company, it seems also to distinguish between the company and its suppliers. The distinction between Apple and suppliers is underscored throughout the texts, particularly with the frequently repeated construction, ‘we require’ (for examples, see Table 4). ‘Require’ is one of the most common words—and the most common verb—co-occurring with ‘we.’ This collocation occurs 104 times in our collection of texts. The verb ‘require’ frequently takes the object ‘facility’ or ‘supplier.’ This appears to be formulaic language used in the text and in tables for reporting audit results. The formulation suggests a power differential between Apple and its suppliers where Apple can impose its requirements. This seems to imply a kind of rule enforcer role for Apple, which thereby seems to adopt governance functions (Crane et al. 2008).

A third aspect of Apple’s language that suggests a ‘legalistic’ orientation is its focus on the Apple Supplier Code of Conduct. As shown in Table 2, ‘code’ (#12) is overrepresented in the text when compared with a reference corpus. Interestingly, it appears very frequently in a word combination with the possessive pronoun ‘our’ (see Table 5). Apple explains that its code draws on “internationally recognized standards” related to worker rights and environmental protection (Apple 2012, p. 3). Notably the Apple code also provides a referent for the Chinese NGOs, who criticize the company for allegedly failing to enforce it (FON et al. 2011a, p. 5). However, the NGOs do not rely exclusively on this voluntary code, but refer also to Chinese environmental law as a basis for claims.

In summary, the language used by both Apple and the Chinese NGOs suggests that Apple’s responsibility involves enforcing rules—a type of governance function that seems characteristic of legalistic cosmopolitanism. Specifically the language emphasizes violations, as well as rules and procedures for regulating supplier relationships, and allocates an enforcement role to the brand. For Apple the limits of this governance responsibility are set by its own code, a position the NGOs appear to use strategically when they critique Apple on the basis of code violations.

Greenpeace and Adidas—Examples of More Moralistic Cosmopolitan Language

Below we discuss the results of our keywords and concordance analyses of Adidas’s Sustainability Progress 2011–2013 reports. These texts include not only information about suppliers but also about the firm’s environmental impact. (We discuss this together with other issues of case and text selection in the Rootedness of cosmopolitan strategies in cultural and institutional contexts section.) Three Dirty Laundry reports, published by Greenpeace International, are also studied. These publications received international media attention, in particular the first report published in July 2011. Results are summarized and then presented below. First, we find that there is less concern in these reports with rules and enforcement of standards than in the Apple and Chinese NGO reports. Instead the focus is on goals of positive change. Second, the relationship between Adidas and its suppliers is described in generally collaborative rather than in coercive terms.

First, keywords analysis indicates emphasis on positive change, particularly with regard to the environment. ‘Achieved’ (#13), ‘milestones’ (#15), ‘target’ (#16), and ‘targets,’ (#17) support the aspirational orientation of Adidas’ reports, as does ‘sustainability’ (#8) (see Table 6). However, as in Apple’s reports there is a concern with ‘compliance’ (#9). Although Adidas mentions its ‘Workplace Standards’ document as a reference to rules, this is not a major focus of its report. It should be noted that ‘audits’ is somewhat prominent ranked 20th on the keyword list, but less so than in Apple texts, where ‘audits’ and ‘audit’ rank 8th and 9th. Instead, Adidas’s document highlights positive achievements and the company’s sustainability efforts.

Table 6 Keywords analysis of Adidas’s Sustainability Progress 2011–2013 reportsa
Full size table

Keywords referring to violations, standards or enforcement are not ranked among the top 20 in Greenpeace’s report, which focuses strongly on the environment and various substances that pollute it (Table 7). For Greenpeace, hazardous chemicals are identified as the problem (e.g., ‘hazardous,’ #1; ‘chemicals,’ #2; ‘substances’, #7; and specific chemicals ‘NPEs,’ #4; ‘NP,’ #10; ‘APEs,’ #11) and any use of them—legal or not—is unacceptable. Greenpeace appears concerned less with violations of existing laws or corporate codes, and more with the inadequacy of the codes (Greenpeace 2011a, pp. 62–63). Greenpeace proposes to solve the problem by advocating a new standard of toxic-free production. It is however important to note, that the role that Greenpeace appears to expect the firm to take is more that of a good and responsible citizen, and not that of a rule developer and enforcer. Accordingly, neither Adidas’s nor Greenpeace’s texts seem to emphasize a rule-enforcing role for the company.

Table 7 Keywords analysis of Greenpeace’s Dirty Laundry, Dirty Laundry 2, and Dirty Laundry: Reloaded reports
Full size table

Second, like Apple, Adidas uses the exclusive ‘we’ to refer to itself throughout its documents. Strikingly, ‘our’ is the most highly ranked keyword, indicating its very high level of representation in the text compared with other informative texts (Table 6). However, despite similar keyword rankings for ‘our’ and ‘we’ in the Apple and Adidas texts (5th and 6th and 1st and 5th, respectively), clear differences emerge. For instance, Adidas never uses the very common Apple phrase, ‘we require.’ This is also evident in the use of the word ‘our’ and its word partners. Among word partners for ‘our’, the most common noun is ‘suppliers,’ and ‘our suppliers’ are a frequent object of Adidas action. However, Adidas tends to define its relationship with suppliers in collaborative terms using words such as ‘encourage,’ ‘support,’ ‘check on,’ ‘work with,’ and ‘reach out’ to suppliers—among many other formulations (Table 8).

Table 8 Examples of ‘suppliers’ concordances from analysis of Adidas’s Sustainability Report 2011–2013
Full size table

The language produced by Adidas and Greenpeace seems to bear characteristics of moralistic cosmopolitan discourse. There is less legalistic language, and instead the emphasis is on positive change. Rather than highlight standards and audits, the company seems to describe its own obligations with reference to targets and milestones indicating positive development. This linguistic analysis is supported by Adidas’s acceptance of Greenpeace’s “toxic-free” standard in 2011. However, the firm appeared to limit its responsibility in implementing this standard. By helping to found an industry group to phase out toxic chemicals, Adidas seemed to indicate that this is a stakeholder issue that should be decided on by the firms involved, rather than an issue for civil society to govern.

Discussion

We began with a goal of better understanding discourses about supply chain responsibility. We suggested that defining obligations in cosmopolitan terms may better describe current realities than stakeholder perspectives. In particular we argued that stakeholder approaches are focused on explaining firm decisions about which stakeholders to respond to, and are rooted in a thinking of value maximization. We further argued that the debates about supply chain responsibility are actually less about specific managerial decisions and more about redefining the role of the firm. We then suggested that cosmopolitan approaches may better account for this new dimension of the debate, and conceptualized legalistic and moralistic poles of cosmopolitan discourse. We conceptually connected these two poles to roles with a governmental function and a citizenship function. Although we would expect that firms may use both moralistic and legalistic language in their arguments, we also believe that the language used in their SCR discourse may help illuminate the discursive preferences of a firm. We thus aimed to identify the dominant discourse in firm and NGO texts by applying corpus linguistic techniques.

Two Forms of Cosmopolitan Discourse

We found evidence of cosmopolitan conceptualizations of supply chain responsibility in both company and NGO discourse. Significantly, different models of this discourse appear, and we found support for both ‘legal’ and ‘moral’ types of cosmopolitan discourse. While the former defines responsibility in legalistic terms and is based on rules and their enforcement, the latter instead characterizes responsibility as an obligation to act for positive change.

In legalistic discourse, the focus is on codes of conduct and standards that are informed by cosmopolitan law. A firm’s responsibilities are limited to the obligations placed on it by these forms of voluntary regulation as well as by law. Further, there is a concern with violations of these rules, and the emphasis is on stopping ‘bad’ behavior. Here, the firm appears to adopt a cosmopolitan law enforcer role. This is not to say that a firm using legalistic discourse may not make more moralist claims, as is shown in the quote from Apple’s 2011 SRPR in the Cosmopolitanism: Responsibility Based on Global Citizenship section. That is to say, predominantly legalistic discourse may also contain moralistic claims.

By contrast, the second model is moralistic. In this discourse, the focus is on aspirational goals. The firm’s obligations are not clearly limited, but the firm works proactively toward cosmopolitan goals. Rather than the law enforcer, the firm resembles a moral global citizen. We do not mean to say that rules are unimportant in the moralistic version of discourse, but rather that they are perhaps less prominent. Adidas refers to its code, and Greenpeace advocates the establishment of legal standards for the toxics that they are campaigning against (e.g., Greenpeace 2011a, p. 9; NGO Interview, Beijing, 31 July 2012). However, rules are not the focus, and indeed there is an aspiration to go beyond the standard.

Strategic Use of Cosmopolitan Language

Our second main finding is that companies appear to use cosmopolitan language to build SCR discourse that supports their positions. Substantial use of ‘we’ suggests that the firms’ supplier responsibility documents are not only informative but also promotional, aimed at building support for a particular group and position (Bhatia 2004; Fairclough 2003). Each company, however, seems to adopt a different discursive strategy for promoting itself as a cosmopolitan firm that takes responsibility for SCR issues. In the first version of legalistic cosmopolitanism, Apple’s language suggests a company that takes responsibility by enforcing its code and by requiring suppliers to meet cosmopolitan aspirations. In the second version of moralistic cosmopolitanism, Adidas’s language suggests its status as a good corporate citizen that is setting high goals for itself and attempting to achieve certain positive milestones with regard to sustainability efforts and its supply chain relationships.

Advocating New Roles for the Firm with Cosmopolitan SCR Discourse

We argued that the discourse about SCR is not adequately captured by stakeholder approaches. One aspect of this argument was that stakeholder theory constitutes an organizational ethical theory that assumes value maximization as the key purpose of the firm, and focuses on managing stakeholder relations (Phillips et al. 2003). SCR discourse, however, goes beyond the question of which stakeholders to consider in a firm’s business decisions. It questions the traditional role of the firm as a value-maximizer and suggests that different roles are needed to fully account for SCR. In the context of CSR, Crane et al. (2008) suggested the roles of corporations as citizen and corporations as government.

Our analysis of the SCR discourse of different kinds of organizations (firms and NGOs) has identified two main tendencies in the language which are particularly visible in the vocabularies used. One pattern of language used in SCR discourse appears to be more legalistic and the other more moralistic. These patterns match well with the distinction suggested by Crane et al. (2008): the moralistic discourse connects to firms’ role as citizens that are members of a global community to which they aim to do good. In contrast, legalistic discourse connects to firm’s government role, where firms act as rule developer and rule enforcer. This seems suitable to describe the role of Apple referring to its own code and emphasizing the enforcement of its own code among its suppliers. Similarly, the Chinese NGOs refer to law and expect Apple to assure that its suppliers adhere to Chinese law and that Apple thus takes a role as law enforcer where law enforcement by the authorities may sometimes be incomplete.

As suggested above, these are the tendencies that become apparent when large numbers of words are studied. We observed that organizations make claims using both kinds of language; however, the vocabulary structure approach (Loewenstein et al. 2012) helped us to identify the dominant vocabulary use. It is further interesting to note that Apple combines its legalistic discourse with moralistic claims, e.g., for workers’ rights. However, Apple does not appear to connect this claim to its own role. As discussed in the Cosmopolitanism: Responsibility Based on Global Citizenship section, it seems to contrast the claim with its own understanding of a more governmental role through which it ensures that the suppliers live up to these cosmopolitan ideals.

Similarly, we identify legalistic claims in the discourses of the firm and NGO which we identified as moralistic. However, while Greenpeace calls for stronger legal regulation of chemicals in its Dirty Laundry reports, the expectations of Adidas are better explained as good citizenship that not only adheres to the law but displays concern for the community, whether or not this is legally required. By the same token, Adidas’s joining the “Roadmap to Zero” has a legal component as it supports a normative standard. However, the emphasis seems to be more on its own adherence to standards that go beyond the law, rather than its role in enforcing compliance with the law or with its own rules.

Implications and Conclusion

Managerial Implications

By exploring two cases of firms’ and NGOs’ SCR discourses, we can draw three main implications for managers. First, in line with Ghemawat’s (2011) call for more cosmopolitan corporations, we argue that firms may need to build a more cosmopolitan understanding in order to address the new concerns and new actors that are emerging in a global civil society (Keck and Sikkink 1998; Kaldor 2003). Second, managers should be aware of the tension between stakeholder approaches and cosmopolitan thinking. Third, the cases indicate that firms can expect various types of engagements with different types of NGOs over these issues, and should strategize accordingly.

Integrate Cosmopolitan Views in Non-market Strategy

We argue that the cases above support Ghemawat’s (2011) argument that firms, while generally rooted in their local institutional environments, must develop the capacity for cosmopolitan thinking. This is particularly the case as firms attempt to develop non-market strategies that address constituencies beyond stakeholder relationships. As the examples from China demonstrate, new groups can emerge at any time and in any place, and we suggest that firms with a more cosmopolitan conceptualization of their role and responsibilities as compared to a stakeholder perspective will be better able to overcome institutional distances to engage with others effectively. This lesson may be particularly important—and difficult—for firms that have less experience in dealing with civil society groups. For instance, although Chinese firms have begun to actively embrace CSR practices, business-civil society relations in China are not strong, and firms may not have had substantial involvement with NGOs (Moon and Shen 2010).

Distinguish Two Approaches to Non-market Strategy: Moralistic and Legalistic

In line with Phillips (2003), we argue that it is important to maintain a conceptual distinction between claims made on the basis of a stakeholder identity and those made on the basis of cosmopolitan ideals. This is important for clarifying the terms of interaction with other actors and for anticipating the strategies others might adopt. Actors have a choice when they make claims, and both stakeholder and cosmopolitan frames are potential resources for interactions between firms and NGOs. Further, different applications of the frames seem possible, including a more legalistic form and a contrasting moralistic one. Managers should be aware of these choices and the strategic opportunities they create.

Adopt Non-market Strategy to Specific NGO Characteristics

The legalistic/moralistic distinction is related to a third implication, which is that the framing of claims may open up different opportunities for firm interaction with NGOs. Here a comparison between our two cases is useful. The first case of Apple and the Chinese environmental NGOs is based on a legalistic, rule-focused cosmopolitanism. This seems to have led to partnership and joint efforts to improve environmental auditing procedures. A rather different outcome was seen in the second case of Adidas’s and Greenpeace’s moralistic discourse, where Adidas decided to contribute to an industry-led approach to eliminating toxics in its products. Interestingly, the more moralistic cosmopolitan claims of Greenpeace appear to go hand in hand with the company’s choice to both accept the NGO claim and to implement its response on the industry’s terms. While legalistic cosmopolitanism may imply a role for NGOs in cosmopolitan governance, moralistic cosmopolitanism may simply require good corporate citizenship from firms, perhaps without exactly specifying what form this participation should take.

Implications for Further Research

The Continuum of Moralistic and Legalistic Cosmopolitanism and its Implications

We focused on two cases, moralist firm and NGO discourse and legalist firm and NGO discourse. We also noted that the distinction between moralist and legalist cosmopolitan language describes tendencies rather than strict and exclusive categories. One may thus understand the concept as a continuum between fully moralist and fully legalist poles of language use. It may thus be valuable for future research to more fully explore the space between the two poles. Moreover, it would be interesting to consider cases in which the discourses of NGOs and firms differ. Looking at the organizations we considered, we find examples of more moralistic language where Adidas has placed pressure on a supplier to adhere to local environmental law, an issue that had been raised using more legalistic language by Chinese NGOs in a recent report on water pollution in the textile industry (FON et al. 2012). On the other hand, the NGO Greenpeace, which we identified as using more moralistic discourse in its Dirty Laundry reports, campaigned in 2006 for Apple to take greater responsibility for e-waste by reducing the use of toxic chemicals and by enhancing product “take-back.” Apple, which we found to use more legalistic language, seemed at first not to react to this campaign, but developed its own initiative in 2007 (Edge 2011). These examples may suggest that the language use and an underlying understanding of firms’ role in SCR as more governmental or citizen-like may be connected to firms’ response to claims from different NGOs. This is an area that deserves substantial further investigation.

Rootedness of Cosmopolitan Strategies in Cultural and Institutional Contexts

Our findings suggest that it may be worth further exploring if the choice of cosmopolitan discursive strategies may be guided by the cultural and institutional ‘rootedness’ of organizations. For the most part, firms and their stakeholders remain ‘rooted’ in their home countries and face challenges when trying to adopt a more cosmopolitan mindset (Ghemawat 2011). This result is predicted by institutional theories, which emphasize that the understandings and actions of companies and NGOs are shaped by their local institutional environments. For instance, the influence of national business systems on CSR practices has been identified, and a more collaborative form of interaction with other social actors may be related to the national business system of the company’s home base (Matten and Moon 2008). Apple, as an American firm, is located in a context in which CSR is more voluntaristic and ‘explicit’. By contrast, Adidas and Greenpeace are based in European countries with more ‘implicit’ forms of CSR based on government regulation and formalized obligations between social actors. Perhaps the distinctions we observe between ‘legalistic’ and ‘moralistic’ discourses are related to these different national business contexts.

Evidence for differing discursive strategies between US-based and European firms has been found previously, and text analyses have shown that for instance US firms are more likely to emphasize financial motivations for CSR whereas European firms draw on both financial and sustainability arguments (Hartman et al. 2007). We find additional support for the influence of local contexts on firms’ SCR discourse in the work of Tate et al. (2010). In a content analysis study of 100 global firms’ CSR reports, the authors found different thematic emphases in the reports of firms from different geographical regions. Whereas US-based firms tended to associate social responsibility with the management of risks, in Japan and Western Europe a link is made between social responsibility and community. The authors define this finding in terms of a proposition: “Among firms located in industrialized countries, US firms view responding to the institutional pressure of stakeholder’s sustainability demands as a part of risk management, while European and Japanese firms view it as part of community integration” (Tate et al. 2010, p. 36). Interestingly, this distinction seems to be consistent with our finding of a more ‘legalistic’ style of discourse by an American firm and a more ‘moralistic’ style by a European firm. Whereas management of risk would seem to imply more legalistic language, one might expect community integration to be discussed in more moralistic terms.

In line with den Hond et al. (forthcoming), we also suggest considering further firm- and NGO-specific factors, such as the longer period for which notions of supply chain responsibility have been discussed in the textile sector than in the electronics industry (Overeem 2009), which might contribute to a greater acceptance of cosmopolitan norms. Whereas Adidas has published its sustainability report for 13 years, Apple has produced its SRPR for just 7 years. Moreover, Apple publishes an audit report that focuses on “supplier responsibility”, whereas Adidas’ reports cover sustainability in general. This explains some difference in the keywords. In addition, Adidas’ 2011 report follows Global Reporting Initiative guidelines. While one might argue that some of our findings could be a result of this institutional difference, one could also see them as active choices for a reporting style; choices that reflect the underlying understanding of the firms’ role as exercising governmental or citizen-like functions. This is particularly reasonable for organizational types and ideologies of NGOs (den Hond and de Bakker 2007), where two different types connect well to the role that NGOs attribute to firms in the SCR discourse (Yaziji and Doh 2009): the social movement NGO Greenpeace is more moralistic, while the watchdog NGOs, the Chinese group is more legalistic. We thus suggest that further research may aim to more fully understand this context.

Conclusion

We have found that firms adopt cosmopolitan language to address NGO claims about supply chain responsibility, and we have considered how legalistic and moralistic cosmopolitan approaches suggest different roles for firms with regard to SCR. While the former describes a role as rule developer and enforcer and thus suggests a more governmental role for the firm in SCR, the latter emphasizes the obligation to do good to the global community and thus promotes a kind of citizen role for the firm. We have also considered two cases in which these different languages could be observed in firm discourse, and we suggested connecting them with different roles for firms in SCR. By questioning the role of the firm and extending it beyond the principle of value maximization, SCR discourse may thus contribute to debates about the legitimization of the firm in society that started from corporate scandals at the beginning of the millennium and increasingly calls for alternative models of firm responsibility.