Introduction

Although Johanson and Vahlne’s (1977) and Johanson and Wiedersheim-Paul’s (1975) seminal contributions to the internationalization process theory (IPT) literature have had widespread influence on the international business (IB) field and beyond, scholars still contest their core insights. Indeed, Håkanson and Kappen (2017: 1103) recently challenged IPT’s core theoretical value, arguing that “the inductive theoretical arguments presented in Johanson and Wiedersheim-Paul (1975) and Johanson and Vahlne (1977) only very partially reflect the empirical observations on which they were based.” The relevance of traditional stage theories explaining firm internationalization (Vernon, 1966; Bilkey & Tesar, 1977; Reid, 1981; Johanson & Wiedersheim-Paul, 1975; Johanson & Vahlne, 1977) has been questioned on several grounds, including that they lack explanatory power (Andersen, 1993; Oviatt & McDougall, 1997), that firms do not or should not follow the stages in the presented sequence (Reid, 1983; Turnbull, 1987; van de Ven, 1992), and that the models do not inform the initiation of the internationalization process1 (Andersen, 1993). Thus, stage models tend to describe the internationalization processes as highly structured and deterministic, even though frequent but unforeseen environmental interactions, and the variable organizational responses to them, suggest that they are anything but (Stubbart & Smalley, 1999).

In response, IB scholars have pursued a number of extensions and alternatives, including those informed by transaction cost economics (Hennart, 1982), internalization (Buckley & Casson, 1976), and neo-institutional theory (Kostova, 1999). Yet each of these perspectives suffers from some of the same shortcomings as IPT due to their largely deterministic assumptions. The somewhat inconsistent findings of IPT and subsequent IB theories likely stem from their failure to incorporate underlying managerial processes that lead to internationalization decisions. In support of this position, Johanson and Vahlne (1977: 23) understood the deficiencies of their model from the onset: “Because we…disregard the decision style of the decision-maker himself, and, to a certain extent, the specific properties of the various decision situations, our model has only limited predictive value.”

In order to fully understand internationalization, theory must elucidate both its content and process (Barnett & Carroll, 1995). We believe a discourse-based view of internationalization could complement traditional IPT by filling in what Vahlne and Johanson (2017) referred to as the “black box” (i.e., dynamic processes). Specifically, we conceptualize a model that emphasizes the embeddedness of organizational discourse in social, political, and historical contexts, and that allows us to trace the dialectical processes that motivate the decision to internationalize (Vaara & Lamberg, 2016). Rather than identifying and documenting internationalization outcomes, discourse analysis allows for the integration of managerial processes through which organizations evolve and are sustained (Phillips & Oswick, 2012).

In general, discourse refers to the practice of writing and talking (Woodila, 1998). Discourse has been defined as a compilation of interrelated texts or “practices that systematically form the objects of which they speak” (Foucault, 1972: 49). The focus is less on the specifics of language itself, “but more on the coherence of the underlying concepts and ideas contained in a particular set of texts and their evolution through time” (Phillips & Oswick, 2012: 10). Collectively, these texts and practices provide a coherent means by which to embody people, events, ideas, and phenomena in a given social context, validating some of these embodiments and ruling out others (Phillips, Lawrence, & Hardy, 2004). Discourse analysis “therefore involves analysis of collections of texts, the ways they are made meaningful through their links to other texts, the ways in which they draw on different discourses, how and to whom they are disseminated, the methods of their production, and the manner in which they are received and consumed” (Phillips et al., 2004: 636).

Discourse-based theorizing has been underutilized in IPT and the IB field, despite language being fundamental to the construction of social reality that characterizes organizational phenomena (Phillips & Hardy, 2002), including in multinational enterprises (MNEs) (Luo & Shenkar, 2006; Brannen, Piekkari, & Tietze, 2014). In this paper, we argue that insights from IPT and other IB theories can be enhanced by dialectical theorizing (Foucault, 1980; Van de Ven & Poole, 1995) and discourse-based analytical approaches (Fairclough, 1995), leading to a dynamic and more complete model of internationalization that integrates both content and process. Building on Foucault (1980), Fairclough (1992), and Van de Ven and Poole (1995), we argue that internationalization processes are grounded in a dialogical conflict, emphasizing the role of knowledge, power, and resistance, ultimately leading to consensus and performativity (see Figure 1). This process is driven by individuals and groups advocating for positions, and is resolved through contestation, managerial processes that IPT and other IB theories have not fully accounted for. Our model allows for the examination of how power and conflict lead to an internationalization decision through the examination of talk and text.

Figure 1
figure 1

A discourse-based perspective on internationalization.

Full size image

We present our arguments in five sections. First, we briefly summarize IPT and review its strengths and weaknesses. Next, we provide a theoretical overview of the organizational discourse literature, highlighting its relationship to the internationalization process. We then develop our own multi-level conceptual model of the internationalization process, integrating insights from the discourse literature and dialectical theory, while linking organizational processes with the contextual environment. We illustrate our model through two contrasting cases of internationalization: Bulova’s successful international expansion, and Google’s entry, exit, and thwarted re-entry into China. Next, we discuss how our perspective can be integrated into Johanson and Vahlne (1977, 2009) and Vahlne and Johanson (2017) to build a more complete model of the internationalization process, inform related IB theories, and explicate new global phenomena, while providing specific suggestions for applying our approach. Finally, we summarize the theoretical contribution of our research to IPT and IB, and illustrate how a dialectical perspective could enable scholars to advance knowledge and theory.

The Internationalization Process Model and Its Criticisms

To explain the internationalization decision and other IB phenomena, scholars have incorporated theories drawn from allied social sciences and introduced unique theories. IPT is among the most original contributions to IB literature, theorizing that firms enter countries where they have more knowledge and experience in a sequential or staged fashion. Both the formal “stage” model (Johanson & Wiedersheim-Paul, 1975) and the nuanced “process” model (Johanson & Vahlne, 1977) draw from the behavioral theory of the firm (Cyert & March, 1963; Aharoni, 1966) and the theory of firm growth (Penrose, 1959). Both rely on the assumption that, with increased market knowledge and experience, firms overcome physical distance and increase their commitments within and across markets.

Recent research has challenged this theory on several grounds, including that there is little empirical support for IPT (Andersen, 1993; Turnbull, 1987; Sullivan & Bauerschmidt, 1990) and that it is applicable primarily to the early stages of internationalization (Forsgren, 1989). Others have argued that some firms can internationalize much sooner than the process model anticipates, effectively circumventing stages (Oviat & McDougal, 1994). Combining these two critiques, it appears that increased international trade, growth of globally integrated supply chains, and an upturn in the digital transmission of value has broadened the universe of firms that can connect rapidly to international markets, reducing the potential pool of companies subject to the slow, sequential process that Johanson and Vahlne (1977) theorized. More broadly, a fundamental criticism of IPT is that it is deterministic and reductionist, and therefore tends to overlook the heterogeneity embedded in industries, firms, markets, and idiosyncratic contexts (Sullivan & Bauerschmidt, 1990).

Another disadvantage is that most IPT studies use measures that are calculated by rough proxies, attempting to impart meaning and motivation into ex post observations without uncovering the underlying decision-making processes that led to them. As such, there is a potential disconnect between the theoretical notion of process and its development and measurement. Despite these widespread critiques, no viable complementary theory of internationalization has been proposed that accounts for underlying managerial processes. This gap between theory and analysis prompted Andersen (1993) to argue that empirical design should be adapted to the theoretical model through a different and more tailored operationalization of the internationalization process itself.

More recently, Forsgren (2016: p. 1142) argued that “If we assume that we need a complete model dealing with the same issues as Johanson and Vahlne’s (1977) model — pace, direction, and commitments of firms’ foreign operations,” then Johanson and Vahlne’s (2009) “revisited model might, at best, provide a fruitful starting point for developing such a model.” Toward that end, in a 40th anniversary update to their original model, Vahlne and Johanson (2017) sought to supplement it by adding capability-creating processes, but the issue of underlying managerial processes remains unresolved (Coviello, Kano, and Liesch, 2017). Supporting the development of a more complete model, Coviello et al. (2017: 1156) argued that “actions of individuals are tightly intertwined with firm-level outcomes, and, as argued by Kano and Verbeke (2015), they should constitute a key ingredient of any credible, managerially relevant theory.”

While IPT and many subsequent IB theories fail to account for the firm’s embeddedness in unique social (Fiss & Hirsch, 2005), political (Melin, 1992), and historical (Vaara & Lamberg, 2016) contexts, IPT was never intended to account for underlying processes in a modern and complex internationalizing firm. According to Welch, Nummela, and Liesch (2016: 786): “Johanson and Vahlne (1977) also caution that theirs is a ‘partial’ model, and they are explicit about its scope. They ‘do not deal explicitly with the individual decision-maker’ (Johanson & Vahlne 1977, p. 26) or variations in decision-making styles (Johanson & Vahlne 1977, p. 23), nor do they consider the conditions that prevail in specific ‘decision situations’ (Johanson & Vahlne 1977, p. 23).” We argue that this dynamic interaction between decision-makers and decision contexts is at the heart of internationalization processes and is imperative to our research.

Building a Discourse-Based Model of the Internationalization Process

We ground our model of internationalization in dialectical theory (Benson, 1977), and support it by leveraging Van de Ven and Poole’s (1995) classification of theories of organizational change processes: life cycle (stages), evolutionary, teleological, and dialectical. Although Vahlne and Johanson (2017) re-characterized their perspective according to evolutionary theory, we concur with Santangelo and Meyer (2017) that their approach is more closely aligned with the teleological tradition because their unit of analysis is a single organizational entity. We argue that this limits its ability to examine the internationalizing firm’s historical processes of learning, adaptation, and change (van de Ven & Poole, 1995) by making interactions between organizations and the environment secondary to the individual organization (Van de Ven & Poole, 1995). However, dialectical theory can complement traditional life cycle and teleological approaches because our unit of analysis is the group of individuals involved in the internationalization decision in relation to the external environment. As such, a dialectical approach allows for the study of interactions between multiple organizations in order to develop a more fine-grained understanding of the processes of competition, cooperation, conflict, and other forms of interaction (van de Ven & Poole, 1995).

According to dialectical theory, organizational stability and change are the result of opposing forces vying for power (van de Ven & Poole, 1995; Hardy & Maguire, 2016) in which actors use discourse to communicate their positions. These positions “then form patterned redundancies in social interaction that eventually get institutionalized into organizational rules, roles, and systems” (Fairhurst & Uhl-Bien, 2012: 1045). These opposing forces may operate internally, as between inter-organizational interest groups (Boje, Oswick, & Ford, 2004), or externally, as the organization’s interests intersect with other organizations.

Our model builds on Phillips and Hardy (2002) and Phillips et al. (2004) by assuming that there is a mutually constitutive relationship between discourse, text, and performativity. More specifically, Phillips and Hardy (2002) argued that discourses are produced and their meaning is shared and consumed, potentially leading to an organizational shift. In developing our model, we adopt Parker’s (1992: 5) definition of discourse as “a system of statements which construct an object” and we employ a performative (e.g., discourse as doing) dynamic perspective (Potter & Wetherell, 1987; Maguire & Hardy, 2013). Building on the idea that scholars should be able to point to an observable resource commitment that signifies an international decision (Oviatt & McDougall, 1994), our discourse-based model makes explicit the relationship between discourse, managerial processes (Morgan, 1980; Maguire & Hardy, 2013), and performativity (Spicer, Alvesson, & Kärreman, 2009).

Guiding Assumptions and Levels of Analysis

We argue that a discourse-based perspective of internationalization must be grounded in four guiding assumptions that are capable of supporting a wide range of theoretical and empirical analyses. First, the social reality that emerges from discourse is not a fixed state but a dynamic process, occurring rather than existing (Sztompka, 1991). This assumption addresses Coviello et al.’s (2017) argument that, because Vahlne and Johanson (2017) ignored individual relationships, their model is static. Second, discourse analysis involves the search for holistic rather than linear explanations of processes (e.g., Santangelo & Meyer, 2017). Third, discourse analysis must link process analysis to the location and explanation of internationalization. And fourth, questions surrounding internationalization should be framed in the language of becoming (i.e., dynamic processes) rather than being (i.e., static outcomes) (Sztompka, 1991), thus demanding detailed, comparative, and longitudinal textual data analysis (Pettigrew, 1992). In order to develop theory that complements Johanson and Vahlne (1977) and Vahlne and Johanson (2017), we agree with Pettigrew (1992: 6) “that in the conduct of strategy process research, the what and the how, the content and the process, are best regarded as inseparable.”

Our model is composed of two interrelated levels: a surface level of communicative actions and a deeper level of discursive structures that are linked recursively via actors’ interpretations of texts (Heracleous & Barrett (2001). Heracleous and Barrett (2001: 758) defined deep structures as “relatively stable, largely implicit, and continually recurring processes and patterns that underlie and guide surface, observable events and actions.” We argue that deep-level structures most often act as a constraint, reinforcing the status quo. However, if they become misaligned with strategy and eventually reconfigured, they can serve as a facilitating mechanism to support an internationalization initiative (Phillips et al., 2004).

Intertextuality in a Discourse-Based Internationalization Model

According to Fairclough (2003), there are three distinct elements in the process of meaning making: the production, dissemination, and reception of the text. However, it is the interplay between them (i.e., intertextuality) that leads to meaning and outcomes. We adopt Fairclough’s (2003: 218) definition of intertextuality (extending Foucault, 1972 and derived from Bakhtin, 1981): “The intertextuality of a text is the presence within it of other texts (and therefore potentially other voices than the author’s own) which may be related (i.e., dialogued with, assumed, rejected, etc.) in various ways.” Thus, when the speech or writing of an individual outside of the primary text is reported, two or more different texts and two or more different voices are brought into dialogue, resulting in potentially two or more different perspectives, objectives, and interests (Volosinov, 1973).

A Discourse-Based Model of Internationalization

We propose that a dynamic internationalization process unfolds in an environment characterized by consensus, dialogue, and the convergence of positions — arrived at with knowledge regarding the firm’s global strategic presence — leading to the accumulation and deployment of power and influence in the face of resistance. We theorize that these positions emerge at the interface of overlapping discursive fields within and outside the firm, and later as discourse within the organization, producing a dominant text that acts as a catalyst to either support or resist internationalization (Hardy & Thomas, 2014; Vaara & Tienari, 2011; Vaara, Kleymann, & Seristö, 2004). A discursive field can be understood as “a symbolic space…that establishes the limits of discussion and defines the range of problems that can be addressed” (Wuthnow, 1989: 13). Once a misalignment between external sociopolitical/economic, environmental, and internal organizational discursive fields has materialized, our model demonstrates how underlying discursive processes, leading to the onset of internationalization, eventually shift the organization back into alignment with its external environment (Poole & Van de Ven, 1989).

Our model has three levels: contextual (environment), dialogical (process), and textual (content). We support our theoretical arguments with two examples of underlying dynamic processes: Bulova’s international expansion in the early 2000s, and the longstanding saga surrounding Google’s decision about whether and how to operate in China.

By the mid-2000s, Bulova had established a strong presence in the North American marketplace as a mid-priced, traditional watch line. During that period, Bulova senior management recognized that the North American market was mature and, in order to grow, Bulova needed to consider international expansion (Perry, personal communication, 2020)2. Despite some naysayers who argued that Bulova should focus on the domestic market, the company established an administrative office in Switzerland to oversee development of a higher-end line using Swiss movements. Bulova also engaged Dennis Perry, a consumer products executive with substantial international experience, as chief strategy officer to assist with the international expansion. Perry concluded that Bulova’s internationalization effort was suboptimal because (1) it created a separate, distinct brand that did not leverage Bulova’s brand image, (2) it relied solely on distributors versus wholly-owned subsidiaries, and (3) the investment was too modest to achieve substantial growth. Instead, he argued that Bulova should pursue a “one brand, one line, one image, one position” strategy. As Perry began to promote the global expansion plan, Bulova’s owners elected to sell the company (Perry, personal communication, 2020). In January 2008, Citizen bought Bulova but committed to allowing Bulova to continue to operate as a separate, autonomous organization. Citizen supported the development of technologies unique to Bulova that would allow it to maintain and strengthen its global brand, and committed not to appropriate any of Bulova’s intellectual property. Although sales dropped substantially in the aftermath of the 2008–2009 financial crisis, Perry helped put Bulova in a position to rebound following the crisis through the introduction of new products, under the name Precisionist, that had a higher frequency crystal and sweeping second hand, which allowed it to increase its price point and continue Bulova’s aggressive international expansion (Perry, personal communication, 2020).

Google’s involvement with China began in earnest in 2007 when, shortly after Google.cn received its license from the Chinese government, the company signed a set of guidelines committing to compliance with censorship (Fannin, 2010). After several years of criticism from human rights activists within and outside of China, on January 13, 2010, Google released a statement saying it was ready to end censorship of its search service. However, CEO Eric Schmidt appeared to back away from this commitment, indicating that Google was planning to stay in China even if it was forced to close down its local search services (Waters, 2010). In March 2010, Google formally announced that all Google.cn users would be directed to the uncensored Google.com.hk website, allowing mainland users access to their search engine. Despite no longer providing Google.cn to China, censorship battles continued. In 2014, China restricted access to nearly all of Google’s auxiliary services, resulting in a drop in market share to less than two percent. Apparently influenced by China’s large and growing consumer market, Google reconsidered its effective withdrawal from China. In August 2018, online magazine The Intercept released a leaked internal Google memo highlighting a new attempt by the company to create a search engine, code-named “Project Dragonfly,” for the Chinese market (Gallagher, 2018). It suggested that, rather than standing by its commitment to provide an uncensored search engine, Google was exploring the possibility of creating a product that complied with the Chinese government’s demands. This led to ongoing discourse between Google and its stakeholders, including employees, regarding whether Google should engage in China and, if so, under what conditions.

Contextual Level (Environment)

Although firms must continually adapt to changes in their external environment, research suggests they may also co-create their environments (Cantwell, Dunning, & Lundan, 2010). Although both levels develop independently, change in the external context can act as a catalyst, prompting an internal response. The process may also be ignited from within, such as when an organizational problem motivates a search of the external environment for an international solution. In either case, informational discourse from the contextual level may overlap with discourse originating inside the organization if it emerges in a time or place close to the firm’s internal challenges or facilitates the firm’s ability to respond to them. Consistent with our view, “Context is the part of a text or statement that surrounds a particular word or passage and influences its meaning, or the circumstances or setting in which an event occurs” (Sillince, 2007: 365). In the case of the internationalizing firm, we argue that the contextual level is composed of sociopolitical and economic cues that may lead to the identification of patterns of meaning (Weick, 1999), but only if these cues are recognized and paired with organizational challenges. Although research has already established that broader access to information, knowledge, and technology through networks of relationships is important to the internationalization process (Johanson & Vahlne, 2009; Oviatt & McDougall, 1994), our objective is to illustrate how individual managers use this external knowledge to initiate an internationalization decision.

In the case of Bulova’s internationalization, Perry initiated a campaign to cut costs, develop innovative product offerings, and prepare for a more aggressive internationalization strategy. In Google’s struggle with whether to remain, withdraw, and later re-enter China, two opposing forces exerted pressure on the firm. Although the commercial potential of a search engine in China was evident, China’s willingness to censor searches and force internet firms to cooperate in periodic crackdowns on political dissidents generated a strong countervailing force, prompting Google to initially withdraw search services from mainland China (Fannin, 2010).

Our model theorizes that, when discourse surrounding changes in the firm’s external or internal environment becomes intense enough, organizational discourse will overlap with discourse in the contextual field, illuminating misalignment that predisposes the firm to a response. Our contextual approach builds upon Phillips, Sewell, and Jaynes (2008: 776), who argued that discourse in organizations “can be studied as cases of intersubjective meaning making that use discursive resources from outside the organization to achieve consensus around its strategic ends and the means adopted to achieve those ends.”

Proposition 1:

In the context of internationalization decision-making, internal or external stimuli highlighting an international challenge or opportunity are likely to be assimilated into internal discourse when those stimuli are intense and out of alignment with the organization’s current internationalization trajectory, and when internal actors identify the potential to develop discourse that (partially) resolves that misalignment.

We view the growing Chinese market throughout the 2010s, coupled with Google’s modest market share there, as reflective of misalignment between external forces and internal consensus. The decision to launch Project Dragonfly and the discourse justifying that decision, as represented in the transcript in Appendix 1a, represents an effort by internal actors to resolve that misalignment. Regarding Bulova, a similar misalignment between growth in global versus domestic markets provided an external stimulus. However, according to Perry, a strong internal effort was put in place in order to mobilize the organization. This push took the form of frequent communications among the top management team, where senior leadership communicated its support of the international expansion plan. These communications were followed by two-way discourse between Perry and organizational members in functional areas whose support was important (Perry, personal communication, 2020). Finally, Citizen made it clear that it was committed to Bulova’s success in global markets, providing additional impetus to the internationalization initiative.

Dialogical Level (Process)

If organizational change is the result of opposing forces vying for power (Foucault, 1980; van de Ven & Poole, 1995), then it follows that consensus, or the absence of dialogue, precedes organizational change. An important contrast between intertextuality and consensus is that the former generally encourages different viewpoints and dialogue while the latter emphasizes the status quo. This suggests that if organizations are designed and sustained through the production, distribution, and consumption of texts (Hardy, Lawrence, & Grant, 2005), then the logical starting point in a critical discourse analysis (CDA) (Fairclough, 1995) of the internationalization process is the absence of change-related dialogue.

To capture the entire internationalization decision process, we conceptualize three levels of texts: informational texts, power-infused texts, and dominant texts. An informational text represents consensus that a company has an internal problem or opportunity that has not been linked with a potential international solution. A second informational text may follow that represents the initiation of discussion (dialogue) about a potential international solution. Next, one or two power-infused texts materialize that support or resist internationalization. Finally, dialogue is resolved and a dominant text emerges that makes the decision about investing internationally clear (consensus). We suggest that cues circulating in the external environment and organizational challenges act as informational texts, as they are not yet infused with knowledge or power. In our model, we consider information as a flow of discourse that might add to, restructure, or change knowledge (Machlup, 1983), and we conceptualize knowledge as intertextualized discourse that has become “a personal ‘belief’” (Nonaka, 1994: 15). The intersection of two informational texts — one reporting conditions or changes in the external environment and the other identifying a problem or opportunity in the firm — constitutes the initiation of dialogue.3 The overlapping of these discursive fields can lead to personal knowledge, when actors inside the firm pair cues from the contextual environment with firm-level challenges that could be resolved internationally.

Because the first step in this process is to specify when and how these two discursive levels interact (Poole & Van de Ven, 1989), we argue that the internationalization process begins when organizational actors acquire knowledge as to how the intersection of the organization and its environment can be used to the firm’s advantage (Fiss & Hirsch, 2005). This newly acquired personal knowledge can then be used to strengthen an actor’s power, as knowledge allows actors to use discourse to enable (or resist) strategy (Vaara & Tienari, 2011; Paroutis & Heracleous, 2013). Organizational actors who developed this personal knowledge resource will initiate the production and dissemination of a power-infused text to advance an international solution to the organizational challenge.

This power-infused text can be guided by sensemaking, as organizational actors are expected to “develop some sort of sense regarding what they are up against, what their own position is relative to what they sense, and what they need to do” (Weick, 1999: 42). From this perspective, organizational actors creatively arrange and reassemble cues obtained from the external and/or internal environments, and these cues provide structure and guidance in a recursive process of organizational enactment (Weick, 1995). According to Fairclough (2003), this is the point when consensus transitions into dialogue, based upon opposing viewpoints vying for power. In the Google case, senior managers appear to have been influenced by knowledge gleaned from external sources that the company could not be successful in the long run without a strong presence in China. These “validity” claims strengthen proposals through the process of dialogical learning.

Our conceptualization of power is consistent with Baldridge’s (1971) coalitional model that emphasized power and conflict and de-emphasized the importance of bureaucratic universal rules and procedures. Thus, when an individual advances an international solution to an organizational problem through the construction of a power-infused text, that influence is strengthened by personal knowledge (Foucault, 1980). Therefore, the objective of power-infused texts is to produce powerful messages that support convergence and lead to widely shared positions (Phillips et al., 2004), culminating in a dominant discourse.

However, Hardy and Thomas (2014: 31) suggest that “the power effects of particular discourses are neither automatic nor deterministic,” because “the discourse has to be intensified through material and discursive practices that normalize and extend its reach.” However, even this may not be enough, as potential obstacles to an internationalization text becoming dominant stem from embedded deep-level structures that have yet to be co-opted (Heracleous & Barrett, 2001).4 Thus, individual or group resistance is likely (Erkama & Vaara, 2010; Hardy & Phillips, 2004), leading other actors to produce a competing power-infused text of their own. Although past empirical research tends to view resistance cognitively, emotionally, or behaviorally (Piderit, 2000), we conceptualize it discursively (Erkama & Vaara, 2010; Hardy & Phillips, 2004).

Following Laclau (1996), we envision that consensus building leading to internationalization stems from negotiation, consolidation of power, and the attenuation of dialogue. If the internationalization discourse becomes a dominant text, the individuals who initially proposed this solution may now possess discursive legitimacy (Fairclough, 1992; Parker, 1992; Phillips & Hardy, 1997), enabling them to influence the social reconstruction of the firm’s deep-level structures, even without formal authority to do so (Hardy & Phillips, 1998). This is because the consolidation of power that results from a definitive internationalization commitment can lead to shifting organizational ideologies (Eagleton, 1991; Van Dijk, 1998) and deep-level structures.5

Proposition 2:

In the context of internationalization decision-making, where informational texts are transformed into power-infused texts that overcome resistance and proliferate through the influence of knowledge-dominant texts, there is an increased likelihood that discourse affects deep-level structures.

We assume that there were additional communications between Google senior management in support of the re-entry decision and that the statement in “Appendix 1a” was shared and reinforced by other senior managers. As these textual representations proliferated, deep-level structures were disturbed, resulting in resistance. Indeed, a group of nearly 1500 employees signed a letter to senior management expressing their resistance to re-entry (“Appendix 1b”). According to the New York Times, this “internal activism presents another obstacle for Google’s potential return to China” (Yuan & Wakabayashi, 2018b). This reflects individual or group resistance (Erkama & Vaara, 2010; Hardy & Phillips, 2004) in the form of a competing power-infused text that reflects their own version of knowledge and power.

Dennis Perry’s strategy to reposition Bulova’s internationalization strategy to be more aggressive is consistent with our coalition model of power and resistance. Perry emphasized that those who were unsure about the earlier internationalization program quickly came on board for the subsequent effort for two reasons. First, Citizen provided “a wealth of information” related to technological expertise and their knowledge of global markets (Perry, personal communication, 2020). Second, Perry took a coalitional approach to power by not demanding action from those he would need to implement his strategy but instead helping them to understand “the role that they could play by being part owners of the plan.” Further, by enlisting Richard Branson as Global Brand Ambassador, Bulova signaled to internal and external constituencies that it was committed to global expansion.

Textual Level (Content)

At the textual level, we focus on the manner in which outcome texts, stemming from discourse, produce meaning in the internationalizing organization.

In our model, texts take on two roles. First, as described at the dialogical level, texts facilitate the study of the social construction of the organization processually, because discourse can only be studied indirectly by examining the texts that constitute it (Fairclough, 1992; Parker, 1992). Second, discourse leads to a dominant text that establishes consensus in advance of performativity surrounding the process of internationalization.6 Seen in this light, texts refer not only to written transcriptions but also to “any kind of symbolic expression requiring a physical medium and permitting of permanent storage” (Taylor & Van Every, 1993: 109). Discourse, on the other hand, can be thought of as structured collections of meaningful texts (Parker, 1992). Thus, “texts can be considered to be a manifestation of discourse and the distinctive unit…on which the researcher focuses” (Grant & Hardy, 2004: 6). Our CDA approach combines the dialogical level (i.e., process) with the textual level (i.e., content) to reveal a more complete model of the internationalization process.

We theorize that the motivation and rationale for internationalization depends on the discursive production of two independent (i.e., not infused with knowledge or power) informational texts. We argue that one informational text stems from a change or previously unnoticed condition in the sociopolitical/economic environment that may be beneficial to the organization. The second informational text originates independently as a problem or opportunity within the firm. Each of these texts remains informational until someone recognizes the advantages of linking them intertextually to produce a power-infused text.

To frame a power-infused text, actors typically articulate their position and strengthen it by pairing knowledge with power connoted by the actor’s endorsement. Separate actors usually produce another power-infused text resisting internationalization, articulating their own version of knowledge and power.

Although there may be hundreds of texts produced and disseminated since the organization shifted from consensus to dialogue, a dominant text can be inferred when a dialogical state shifts back to consensus (i.e., absence of dialogue). Although a researcher will be able to identify and access a dominant discourse, the challenge is not only to collect a sufficient body of preceding texts but also to narrow the focus in order to ascertain the logic (i.e., intertextuality) behind the shift. In practice, a dominant text can be the international solution as originally proposed, an alternate negotiated international solution, or no action.

Proposition 3a:

In the context of internationalization decision-making, when resistance persists and deep-level structures are disturbed but not realigned, there is greater likelihood of no action or partial (intermediate) action.

Proposition 3b:

In the context of internationalization decision-making, when a dominant text leads to consensus and a reconfiguration and realignment of deep-level structures, there is greater likelihood of performativity (action).

During Bulova’s internationalization effort, the resistance centered on the question “What is the expanded internationalization going to mean for my interests?” In addition, there was a much stronger realignment of Bulova’s deep-level structures following the Citizen acquisition (Perry, personal communication, 2020). The main difference in the two phases was the emphasis that Citizen placed on Bulova’s global outlook. According to Perry, the most fundamental change in Bulova’s deep-level structures was that Bulova’s U.S. headquarters went from being “historically a source of support for the U.S. marketplace to one of central support for global operations.” In practice, functional areas, such as marketing, began to envision their job as a “global support network” (Perry, personal communication, 2020).

At Google, the internal opposition to re-entry persisted as Google employees communicated their formal opposition to the Dragonfly initiative (“Appendix 1b”). It appears that the interaction of these positions, as reflected in formal texts, resulted in continued resistance and a new dominant discourse that supported a retreat from the Dragonfly initiative while reserving the potential for future limited engagement. While Google made a US$550 million investment in JD.com, China’s largest online retailer, and established a massive research center in Shanghai focusing on artificial intelligence, it abandoned plans for a more significant re-entry.

Figure 1 summarizes our model by depicting the initiation of discourse via the interaction of external stimuli and the production of informational texts,the infusion of knowledge and power leading to power-infused texts; the role of resistance in constraining, delaying, or limiting the development of a dominant text, and the potential for that dominant text to (1) disturb deep-level structures but not performativity because of continued resistance, or (2) realign deep-level structures leading to performativity (action).

Discussion and Contribution to IB Theory and Phenomena

Dialectical theory and discourse methods can strengthen and complement IPT to reveal processes undergirding internationalization decisions and contribute to other IB theoretical traditions. Further, they can help inform new international phenomena that do not necessarily align with the assumptions underlying IPT or other IB theories. These dialectical and discursive perspectives can complement traditional IB theory, inform emerging international phenomena, and strengthen historical methods in IB.

Insights for Neo-Institutional IB Theorizing

Because IB research using an institutional theory framework has become ubiquitous (Buckley, Doh, & Benishcke, 2017), we accord it special attention. Drawing from Scott (1995) and others, this framework examines how differences in the normative, regulative, and cognitive pillars in the institutional context pose challenges for MNEs, especially in their quest for host-country legitimacy (e.g., Kostova, 1999; Kostova & Zaheer, 1999; Xu & Shenkar, 2002). Although institutional perspectives have explicitly acknowledged the importance of context in IB decisions (Li, Yang, & Yue, 2007; Lu, 2002) extant institutional theory literature does not allow for direct observation of these cognitive behaviors. As such, Kostova, Roth, and Dacin (2009: 997) challenged its validity in informing multinational strategy, arguing that MNEs have a very different “institutional story” than domestic firms that “better fits the conditions of equivocality, ambiguity, and complexity.” They advocate acknowledging that the “social embeddedness of organizations are intertwined in the ideas of agency, social construction, and power and politics” (Kostova et al., 2009: 1003), elements that may be more easily understood using a dialectical approach.

More broadly, by facilitating the study and integration of multi-level phenomena, discourse can benefit theory development in IB’s application of institutional literature. This is because individual actors and their related discourse have been found to reproduce and transform institutional structures (Parker, 1992), yet they have been left out of institutional theory models. Indeed, “the discursive constitution of society does not emanate from a free play of ideas in people’s heads but from a social practice which is firmly rooted in and oriented to real, material, social structures” (Fairclough, 1992: 66). Thus, to gain insight into the effect of discourse on institutions, it is important to analyze the impact of individual discourse on organizational discourse and how interactions among competing firms and/or firms and host country stakeholders result in institutional development. Table 1 compares IPT, neo-institutional, and discourse theorizing as they relate to IB in terms of their core theoretical assumptions and how they treat the constructs of power, knowledge, resistance, and legitimacy.

Table 1 Theoretical approach and treatment of key constructs in IM by internationalization process, neo-institutional, and discursive perspectives
Full size table

In addition, dialectical theory can also act as a powerful tool in other more distant areas, including those that rely on language (Brannen et al., 2014), communication (Ghoshal, Korine, & Szulanski, 1994), and societal differences (Hofstede, 1980) to explain phenomena. The cross-cultural and comparative management literature could benefit from a discursive perspective because cross-cultural research has often focused on individual variables to the exclusion of potentially influential societal norms. Because discourse allows for the exploration and integration of multi-level phenomena, it could facilitate theory development that embeds micro-level discourse in the broader framework of societal communication. Integrating discourse with cross-cultural management could allow researchers to better analyze individual cross-cultural discourse within a broader societal discursive context, rather than relying on individual data as a proxy (Hardy & Phillips, 1999). Further, the process in which knowledge acquisition and dissemination leads to the adoption and legitimation of discourse via the exercise of power could reveal deeper insights into how cross-cultural differences inform managerial actions.

Discourse may also augment the burgeoning international entrepreneurship literature, which has struggled to reconcile micro- and macro-level perspectives that motivate small, medium, and new organizations to become global. This tension between the role of the individual entrepreneur and the influence of the external institutional economic environment has resulted in a disjointed literature, and the emergence of “born global” firms that appear to circumvent the typical staged process further confounds this research. CDA has the potential to reconcile some of these perspectives by integrating and unifying data and evidence of both the individual entrepreneur and the external environment. Specifically, discourse could provide deeper insights into the motivations, deliberations, and communications preceding internationalization, especially when it is rapid and/or coincides with the founding of the firm.

Operationalizing Discursive Approaches: History, Interpretation, and the Democratization of Text

Buckley (2016) advocates the use of historical methods as an approach to incorporate both path dependent/learning influences and environmental factors in firm-level internationalization. He proposes four specific methods: source criticism, time series analysis, comparative methods, and counterfactual analysis. He argues: “Examination of these methods allows us to see internationalization processes as a sequenced set of decisions in time and space, path dependent to some extent but subject to managerial discretion” (Buckley, 2016: 879). Jones (2000) argued that source criticism — the evaluation of historical records including company archives, public statements, and executive interviews — can be among the most effective among historical methods. However, as both Jones (2000) and Gottschalk (1950) observe, this approach assumes that many historical records are absent or incomplete, requiring the researcher to critique sources and interpret the presence or absence of material ex post.

To augment these approaches, studying texts systematically would allow researchers to empirically examine the relationship between discourse and internationalization, because the visibility of texts “provides a focal point for data collection, one that is relatively easy to access and is amenable to systematic analysis” (Phillips et al., 2004: 636). Although textual analysis is an important part of our internationalization model, it ultimately relies on how texts contribute to the emergence and development of discourse (Phillips et al., 2004). However, Barry and Elmes (1997: 430) expressed concerns about the interpretation of texts by asking “Who gets to write and read strategy? How are these acts of reading and writing linked to power? Who is marginalized in the reading/writing process?” Although these concerns remain valid, changes in how written communication is created, disseminated, and recorded may help by offering new strategies for accessing source material. In particular, the ubiquity of electronic communication — both public and private — provides more source material that allows for several research strategies that can improve discourse perspectives. Relatedly, through the increasing use of social media and inter-company email, researchers now have access to communications that, in the past, were mostly limited to internal distribution. For example, key correspondence regarding the internal debate at Google, transcripts of a town hall meeting, and the subsequent online petition were made accessible to the general public. This “democratization of text” provides vast new possibilities for scholars to use CDA in a more systematic way and to reduce the potential for mistaken inference and misattributed meaning.

Integrating these points and drawing from the guidance of Phillips and Oswick (2012), IB is especially suited to the use of multi-level and multi-method approaches that CDA facilitates.

Phillips and Oswick (2012) note that multi-level studies employ CDA by connecting local texts and wider social practices, focusing on organizational events as represented in text, and analysis of pieces of text combined with contextual synthesis. Further, they advocate approaches that combine abstract discursive-narrative approaches with those that recognize the “materiality” of actors and their positions, methods that may fall outside the domain of traditional discourse. One specific illustration relevant to our model and IB research is the cultural-historical activity theory (CHAT) approach, as summarized by Phillips and Oswick (2012: 469), in which discourse is viewed

“as an activity system…driven by shifts, disruptions, and remediations in participants’ engagement with their evolving object (Foot & Groleau, 2011, p. 15). As Foot and Groleau (2011) explain, CHAT-based analysis is “a robust theory grounded in interaction and materiality, that accounts for multiple actors’ perspectives in explaining disruptions and changes as collective practices emerge, coalesce, and evolve (p. 15).

This perspective — the integration of CDA with “material” developments in which discourse is naturally embedded — is just one useful approach to the empirical analysis of IB phenomena through a discursive lens, while also leveraging extant IB theory. Dialectical theorizing and discourse approaches are uniquely suited to resolve some of the shortcomings of IB theories and inform contemporary IB phenomena. Further, drawing inspiration from the call for greater use of historical perspectives in IB, discourse offers several methodological opportunities for explicating the dynamics of internationalization decision-making.

Contributions, Limitations, and Conclusion

Although significant advances have been made since Johanson and Vahlne (1977), little progress has been made on understanding the “complex processes…across time and space” (Pettigrew, Woodman, & Cameron, 2001: 697) that lead to internationalization. Taking a dialectical approach allowed us to develop a multi-level discourse-based model of the internationalization process and to theorize broad processes underlying this transformation. Whereas Johanson and Vahlne (1977) did not emphasize the decision-maker or decision contexts, we respond to the call for alternative theory-building strategies, namely those that “look for theoretical tensions or oppositions and use them to stimulate the development of more encompassing theories” (Poole & van de Ven, 1989: 563). Rather than emphasizing internal consistency throughout the internationalization process, we look first for opposition and tension (i.e., dialogue). We then conceptualize the role that knowledge and power play in producing and disseminating a dominant discourse, and only then the resolution that follows hegemony building, ultimately leading to consensus and performativity. We respond to the call to investigate how individual-level characteristics impact state and change variables surrounding internationalization at the firm-level (Coviello et al., 2017) by attempting to fill in what Vahlne and Johansen (2017: 1089) referred to as a “black box.” Consistent with Phillips et al. (2004), IPT scholars generally focus on patterns of action (i.e., stages). In contrast, we argue that “it is not action, per se, but rather the texts that describe and communicate those actions” that constitute internationalization (Phillips et al., 2004: 635).

Contributions to Internationalization and Other IB Theory

We argue that, in order to understand how and why internationalization is initiated, it is imperative to understand the sociopolitical and economic environment in which an organization is embedded. This is because, in our conceptualization of the internationalization decision, the decision process itself is the focal unit of analysis. When the unit of analysis is embedded in an environment that shapes it and is shaped by it, then this dynamic external environment becomes a necessary part of the investigation (Pettigrew et al., 2001).

We respond to the incomplete scope of IPT and other IB theories as they relate to knowledge, namely IPT’s treatment of experience as a proxy for knowledge and a generic quality comparable across firms. Instead, we side with Tsoukas and Vladimirou (2001), who suggest that knowledge can be appropriated by individuals, but that individual knowledge can also diffuse throughout the organization. Our logic builds on the claim that knowledge is personal (Polanyi, 1962; Nonaka, 1994) but also on Wittgenstein’s (1958) suggestion that knowledge is collective. We add that, seen through a dialectical lens, knowledge is also socially constructed (Foucault, 1980), and that a longitudinal analysis is required to illustrate knowledge as personal, collective, and socially constructed. In our research, we attempt to resolve longstanding questions surrounding how knowledge is certified and shared throughout the organization (Van Dijk, 2006). Our reasoning is consistent with Latour’s (1993) view of knowledge as dynamic and contained within actor networks, acknowledging that knowledge is a resource but also a process.

If personal knowledge can be transformed into organizational knowledge, then logically this process must involve learning. Although traditional internationalization theories view learning through a cognitive lens (Rajagopalan & Spreitzer, 1996), our logic dictates that learning, like knowledge, would be socially constructed. As such, we believe that the process of learning becomes clearer when viewed through a dialogical lens. By reframing the assumptions about knowledge and learning in the internationalizing organization as dialogically constructed, we enable the IB field to develop an understanding of knowledge creation that is socially constructed rather than cognitively understood.

Indeed, the dialogical learning that stems from the influence of validity claims based on the accumulation of knowledge (Habermas, 1984) offers a useful counterpoint and complement to IPT, in that it acknowledges the potential of learning and experience to influence internationalization decisions, but with greater precision and insight as to the actual processes that undergird the action. While IPT assumes that past experience naturally carries over into subsequent decisions, a dialectical approach to learning reveals the dynamics between power and knowledge in validity claims that can influence actions. Further, a discourse perspective directly addresses the question of how individual managerial preference and advocacy can advance an internationalization initiative, including how that advocacy manifests through the exercise of power, the exploitation of knowledge, and the adoption of common positions based on the accumulation of texts, leading to a dominant discourse and performativity.

We build on the theory of structuration (Giddens, 1976) with an eye toward resolving a longstanding paradox: how do individuals maintain control of their own behavior when they are constrained by organizational or institutional structures? We believe that a discourse perspective allows us to correct for this disconnect (Coleman, 1986) by theorizing that, when individual actors possess knowledge and power from pairing organizational challenges with heretofore unseen global solutions, they are in a position to both overcome the constraints imposed by the organization’s deep-level structures and to eventually transform them. Seen in this light, organizational actors who have discursive legitimacy — often through validity claims that have been corroborated — produce discourse that becomes mobilized and disseminated. In this adoption mechanism, power is exercised (Dutton & Duncan, 1987) through the development of a dominant discourse and further manifested in the legitimation of transformed deep-level structures.

Building on Archer (1982), we argue that IB scholars should document the discursive nature of an organization’s deep-level structures: (1) when the organization did not have any international operations and when deep-level structures led to strategies that did not include internationalization; (2) when discourse begins to articulate a major strategic shift that sets the stage for possible deep-level structural change; and (3) during a period of structural transformation in which major changes in deep-level structures become legitimized. This top–down process may occur actively, as organizational actors solidify the gains they made in transforming the organization’s strategy, or it may occur passively, as the implementation of a new international strategy is widely viewed as beneficial.

Limitations

Because internationalization involves a complex interaction of variables, we caution that the application of our model is not likely to be straightforward. In practice, the shift from one organizational mode to another, such as from consensus to dialogue, may not be easy to isolate. Indeed, Mintzberg and Waters (1990) suggested that organizational decisions are neither simple nor straightforward, and that organizational commitment tends to be complex and mostly unobservable. We counter that discourse analysis offers the IB field a viable means to uncover not only the time and the place of an organizational commitment but also the underlying processes that produced it. In order to make better sense of organizational decisions, a dialogical perspective provides researchers with better tools to determine how organizational decisions, such as internationalization, may be understood in advance of performativity.

Our model is best seen as a complement — not a substitute — for other theories of internationalization. Specifically, we bound our model to the processes and dynamics surrounding a specific international opportunity or challenge, and our model ends and other theories (e.g., Vahlne & Johanson, 2017) take over when a decision to internationalize is made. We have omitted explicit examination of how the dialogical process is influenced by prior decisions and experiences, which is the focus of IPT. Those forces are depicted by the dotted lines in Figure 1. However, our model can accommodate those experiences in the form of prior knowledge that actors bring to the discursive process and the dialogical learning that strengthens the development of a dominant discourse. Specifically, past experience will be reflected by — and integrated in — discourse in a way that could reveal precisely how those experiences are interpreted and interact in response to external stimuli. As such, the functional integration of discourse within other IB theoretical traditions can help to better reveal processes leading to an internationalization decision and to more deeply explore insights from more traditional approaches.

We critique our model based on Thorngate’s (1976) postulate of commensurate complexity, which suggests that a theory of social behavior cannot be simultaneously general, accurate, and simple. Ours is a general model of the internationalization process of the firm; it can describe and explain the antecedents and processes leading to international outcomes. Although our model can lead to a more accurate explanation of the origin and processes of internationalization, the tradeoff is the loss of simplicity. This may be necessary, however, given the complexities inherent in dialectical analytics, coupled with the multi-faceted and divergent rationales for contemporary internationalization.

Conclusion

We have sought to offer a genuinely new perspective on the internationalization process and IB theorizing more broadly, one that can lead to a deeper understanding of the motivations behind internationalization decisions and the processes underlying them. Indeed, the objective of discourse analysis is not to accept the social world as it is but to understand “the ways in which the socially produced ideas and objects that populate the world come to be, or are enacted, through discourse” (Phillips & Oswick, 2012: 443). It is our hope that the development of a dialectical approach will facilitate theoretical development in IPT and the resolution of additional unexplained phenomena in IB literature.

Notes

  1. 1

    Consistent with Aharoni (1966), we argue that the decision process is inconsistent with a single, identifiable act, but rather involves a dynamic social process and a succession of acts.

  2. 2

    The description of Bulova’s internationalization journey is drawn largely from a personal communication with Bulova former CEO Dennis Perry.

  3. 3

    We describe in detail the elements that comprise these texts in the subsection below, in which we discuss the textual level of discourse.

  4. 4

    As noted earlier, deep-level structures reflect embedded assumptions, norms, and practices that are taken for granted and widely accepted throughout the organization. Examples include core values, established procedures, and underlying principles. Often the foundation for deep-level structures is established during the firm’s formative years, becoming reinforced and firmly rooted over time.

  5. 5

    Note that these shifting ideologies and the change in deep-level structures may result in a compromise, intermediate position, or another partial outcome, not necessarily a “go” or “no go” decision.

  6. 6

    The progression of our dialectical model ends, or transitions into Vahlne & Johanson’s (2017) model, when a decision is made to invest internationally (i.e., consensus).