Sailing through the storm: Improvising paradox navigation during a pandemic

Paradox navigation, leaders’ learning, and paradox mindset

Many studies have stressed how leaders’ paradoxical understanding of the situation lies at the heart of paradox navigation. Miron-Spektor et al. (2018: 27) emphasized how a paradox mindset that is, the tendency “to value, accept, and feel comfortable with tensions . . . see tensions as opportunities, [and] search for both/and strategies” is key for leaders. Hahn et al. (2014) suggested that the most successful leaders in sustainability attend cognitively to both business and social missions. From this perspective, such a paradox mindset allows leaders to leverage tensions with agility (Pradies et al., 2020), thereby improving their “in-role job performance and innovation” (Miron-Spektor et al., 2018: 26). In sum, being able to think paradoxically is a key skill for leaders to hone, because it enables them to go beyond traditional tensions and embrace novel solutions that reap benefits from opposing poles (Smith, 2014).

Not surprisingly, researchers have been particularly keen on exploring how such a mindset may be acquired. Indeed, welcoming and feeling energized by tensions (Miron-Spektor et al., 2018) is all the more important as contemporary environments become more dynamic and complex. Scholars have noted the progressive efforts that “learning to live with the paradox” (Lewis, 2000: 764) entails. For instance, managers learn to work through paradoxes (Lüscher and Lewis, 2008) as they engage in double-loop learning (Argyris, 1993) and progressively find a link between seemingly contradictory elements. In turn, they develop second-order thinking and pursue paradoxical solutions such as transcendence and synthesis, where either/or assumptions are altered so that opposites can be accommodated (Bednarek et al., 2016). As a result, leaders avoid mild compromises and find ways to optimize oppositional demands (Smith and Lewis, 2011).

Learning to embrace paradoxical tensions takes time, as leaders must integrate lessons from past experience. For example, as time passes, leaders learn to creatively bridge oppositions in their discourses of transcendence (Abdallah et al., 2011). Similarly, organizational actors defend the idea that “the capacity to encompass paradox can be learnt over time” (Bednarek et al., 2016: 13). With time, managers learn “to act in new ways,” thereby avoiding getting stuck in vicious circles and coming to recognize that both sides of a paradox are equally important (Jarzabkowski et al., 2013: 270).

Leaders’ learning to work through paradox, uncertainty, and improvisation

Such progressive learning can be challenged by uncertainties. Indeed, paradoxes are fundamentally full of uncertainties. Paradox scholars not only recognize that the uncertainty of change heightens paradox salience (Smith, 2014), but they also document the intrinsically uncertain nature of paradoxes (Smith and Lewis, 2011). The very “juxtaposition of competing demands creates uncertainty” (Smith and Besharov, 2019: 31), and when centered around core organizational goals produces “organizational instability” (Jay, 2013: 138). Overall, paradoxes “defy rational, linear logic, and in so doing create uncertainty and ambiguity that are both emotionally and cognitively threatening” (Pradies, 2022; Smith, 2014: 1594).

It is thus not surprising that prior research has stressed how leaders work through paradox and its associated uncertainties by progressively reaching a form of “workable certainty” (Lüscher and Lewis, 2008: 228). The literature portrays leaders who learn to live with the uncertainties inherent in paradoxes. For instance, as Lego underwent a radical restructuring, middle managers struggled through a messy process of paradox navigation to achieve a more “workable certainty” that would enable action (Lüscher and Lewis, 2008: 228). Such an effort is also visible in leaders’ actions as they progressively hit their stride while navigating paradoxical situations (Pradies, 2022). Indeed, Smith and Besharov (2019: 25) observed “leaders learning to manage tensions between their dual missions more effectively and efficiently.” The cycle of swinging between opposing poles becomes progressively shorter and less extreme as an organization bounces between the guardrails that it has set, following its own oscillating pace (Smith and Besharov, 2019). Members of a cooperative described balancing pragmatism and idealism as a “challenging task, and one which forces constant change as we grow and learn from our own, and others’ mistakes” (Ashforth and Reingen, 2014: 486). They referred to a quasi “planned uprising” (Ashforth and Reingen, 2014: 500) every few years, following a regular, quite predictable, pendulum movement. Similarly, Jay (2013: 150) described Cambridge Energy Alliance leaders’ efforts to address both public and private mission as a progressive “learning journey” that converged toward the optimization of opposite poles in a synthesis that extended beyond them.

While insightful, such accounts underplay how highly volatile events can call into question the progressive learning journey embraced by leaders. When mentioned, changing exogenous forces in the environment are presented in the literature as creating a window of opportunity that supports internal forces toward reaching a synthesis (Raisch et al., 2018). Overall, while scholars acknowledge the role of the environment in initially surfacing tensions, paradox scholarship does not explicitly focus on the (often fast paced) change in environmental conditions (Smith and Tracey, 2016). Cunha et al. (2020) highlighted an exception, stressing that organizations can appeal to improvisation, defined as “spontaneous action without preparation” (Hadida et al., 2015: 440) when facing unexpected paradoxical tensions: “Effective improvisation becomes especially important the more environments become unpredictable. . . . It can go beyond traditional resolutions of paradox such as segmentation or dialectics to approach sustained synthesis without replacing either element” (Cunha et al., 2020: 3). Yet how improvisational paradox navigation occurs in the midst of a crisis remains unclear. We thus ask: “How do leaders navigate paradoxical tensions when improvising in the face of highly unpredictable and quickly evolving events?”

The narrative lens and temporality

A narrative can be defined as “a semiotic representation of a series of events meaningfully connected in a temporal and causal way” (Onega and García Landa, 2014: 3). For example, organizational actors make sense of unusual experiences by building narratives (Vaara et al., 2016; Vincent-Ponroy et al., 2019). A narrative lens is useful for addressing our research question, as it helps identify paradoxes that are “socially constructed through actors’ rhetoric and conversations” (Lewis, 2000: 771). By inviting reflection on both the past and the future, this lens orients our attention to the temporal dimension of both paradox navigation and learning (Vaara et al., 2016).

Research setting

We answered our research question by examining how French President Macron managed the tension between the paradoxical poles of “saving lives” and “preserving life as usual” through his narratives. As an unpredictable crisis developing at an unexpected pace and for which no government was prepared, the pandemic provided an ideal setting for examining improvisation. Furthermore, it made paradoxical tensions more salient as resources (e.g. time, hospital capacity) became scarce (Carmine et al., 2021; Keller et al., 2021). Finally, given its long duration and multiple phases, the crisis provided the perfect context to examine how paradoxical tensions are navigated over time.

Given the magnitude of the outbreak, political leaders have been the object of much academic scrutiny (Tourish, 2020), with recent work suggesting that a paradox mindset was key (Giustiniano et al., 2020). We focused on the narratives of French President Macron for two reasons. First, Macron’s propensity for paradoxical thinking was visible even before the COVID-19 crisis in his abundant use of paradox management concepts. These also figure in books and articles about him, which we collected using the Europresse and Factiva databases. Examples include his catchphrases such as “yet, at the same time,” his claims to be “both from the right and from the left” during his political campaign, and his explicit use of the words “paradox” and “transcend,” which journalists also emphasized to characterize his way of thinking. Second, France provides an interesting case in terms of timing, with the first wave of virus propagation arriving after Italy and before Germany, theoretically giving France some time to learn from the first case, but not enough to sufficiently plan in advance, thus forcing its leader to improvise.

Data collection

We collected data about all public interventions related to COVID-19, which President Macron implemented between January 2020 (when the first COVID cases in France were declared) and 31 May 2021. Public interventions include official speeches from the presidential palace (n = 8), press conferences and on-site visits (n = 15), spontaneous interviews and more informal forms of expression during field visits (n = 23), and other miscellaneous communications (n = 37) such as posts on social media (e.g. Tweets) or statements recounted indirectly in articles (e.g. a journalist retelling what the president said). The data were collected from various websites including the official sites of the presidency, the “Public life” database, which stores all discourses from French officials, and French newspapers, using keywords related to the pandemic. In addition, we examined President Macron’s social media accounts (Facebook, YouTube, Twitter, Instagram) to develop a holistic view of his narratives across all communication channels (Tuncalp and Lê, 2014). When official transcripts were unavailable, we transcribed the content, which totaled 214 pages. We also read the few academic articles written by Macron himself (Macron, 2000, 2011). Finally, we used additional data (newspaper articles and the official website of the French Health Authority) to establish a chronology about the COVID-19 situation in France.

Data analysis

We first built a chronology of events, paying particular attention to the broad changes in narratives as the contextual situation evolved (e.g. propagation of the virus resulting in Macron’s interventions). Living in France, both authors drew on their personal experiences to make sense of the data, resulting in a rough overview of the possible phases. We then coded Macron’s interventions through an iterative analysis (Deye and Fairhurst, 2019), which allowed us to pinpoint more precisely when the narratives shifted. First, we applied open coding to Macron’s discourse, listing a broad range of topics. While our approach was inductive, prior theoretical concepts can be used in part to guide narrative analysis (Vaara et al., 2016). Thus, in subsequent rounds of coding, we paid particular attention to improvisation, possible paradoxical tensions, as well as the learning related to this improvised paradox navigation. Concretely, we coded all instances of Macron’s discourse involving unplanned responses (e.g. ad hoc actions, expression of surprise), dualities (e.g. economy vs health, open vs closed), or learning (e.g. mentions of past failures, lessons, and comparison with other countries). This enabled us to identify the improvised navigation of the “saving lives” versus “preserving life as usual” paradox central to our study. The paradox is visible through nested tensions (Pradies et al., 2021b), where “saving lives” refers to health-oriented narratives (e.g. lockdowns to prevent the spread of COVID-19), whereas “preserving life as usual” refers to narratives that relate to various facets of normal social life, including cultural life (e.g. education, leisure) or economic considerations (e.g. employment, working conditions).

Between the data and the literature, we then inductively derived three core concepts around paradox navigation. For instance, we originally coded abrupt changes in the trajectory of paradox navigation as “turnarounds.” Later, we found that this idea resonated with the “turning points” associated with the narrative lens. Similarly, the difficulties of foreseeing the dynamics of the pandemic led us to uncover Von Clausewitz’s (1943) concept of “fog of uncertainty,” which directed our attention to hidden dynamics and prompted us to take a closer look at weak signals. As no pre-existing concept (e.g. improvisational or trial-and-error learning) adequately reflected the difficult, unstable, and nonlinear nature of the learning process we were observing, we developed our own label of “chaotic learning.” Finally, we brought these concepts back together, inscribing them in the temporal ordering of the narratives we initially identified into a model of improvised paradox navigation in stormy conditions, illustrated in Figure 1.

OPEN IN VIEWER

Figure 1.

A model of improvised paradox navigation in stormy conditions.

Macron, a leader with a paradox mindset even before the pandemic outbreak

Before the onset of the pandemic, President Macron had become famous in France for embracing a “both/and” view of the world and was viewed by many as a leader with a paradox mindset, that is, one who values tensions and seeks both/and strategies (Miron-Spektor et al., 2018). Such a mindset is evident in his claim to transcend a deep-seated political divide that has been at the heart of the contemporary French political landscape: the left/right opposition. For instance, Macron stressed that his new party would be “both from the right and from the left” (Interview, 13 March 2017). His willingness to engage in both/and thinking is also apparent in his motto “at the same time,” reflecting his commitment to accommodating paradoxical demands: “I said ‘at the same time.’ I was told it’s a verbal tic. I will continue to say it in my sentences and in my thinking, because it means taking into account principles that seemed to be opposed” (Macron, during a political meeting, 17 April 2017).

Journalists and communication experts stressed that Macron’s motto “at the same time” refers to “a resolution of contradictions” that is “representative of a form of government” that seeks “a kind of synthesis” (Mongaillard, 2018). This “language tic makes it possible to link two things that appear logically irreconcilable,” a linguist points out, adding, “The expression gives the impression that synthesis is possible, by refusing to think about the conflict and to arbitrate it” (De Menthon, 2020). Furthermore, although his academic writings are scarce, Macron has offered a vision of political action imbued with a paradox mindset, stating that short- and long-term imperatives within political actions are opposite temporalities that remain deeply interrelated and require both/and actions while calling for the creation of an ideology that helps transcend opposites (Macron, 2011). He notes,

Over the past twenty years, a hiatus has gradually developed between the emergence of long-term, complex, sometimes global, structural problems and economic emergencies whose immediate non-treatment is perceived as unbearable by public opinion [. . .] Political action is thus torn between these two temporalities: the long term, which condemns it . . . and the short term, which calls for imperfect and insufficient urgency. Because the State is expected to manage the immediate, for which there is little to be done, or the very long term, which alone is immortal. The aporia of contemporary political action is precisely linked to the fact that torn between these temporalities, politics does not seem to build a clean and durable action. (Macron, 2011: 107, 110)

Political analysts (Dosse, 2017; Mongin, 2017; Mongaillard, 2018) and Macron himself (Macron and Fottorino, 2017) have commented on this both/and perspective, attributing this mindset to Macron’s proclaimed intellectual filiation with French philosopher Ricoeur, who placed paradox at the center of his thinking (Ricoeur, 1957). Macron himself claimed that Ricoeur “was the person who marked [him] the most, [along] with [his] grandmother” (Mongin, 2017: 26) and he is well-versed in Ricoeur’s writing, as his review of one of Ricoeur’s book signals, which includes multiple references to “paradox” (Macron, 2000: 31). According to some analysts (Dosse, 2017), Ricoeur’s influence is visible in Macron’s willingness to transcend dualities:

Macron’s motto “at the same time” refers to the demands of a tensive thinking that privileges the “and,” as Ricoeur does, seeking mediations to think together and articulate situations of double constraint, irreducible poles, incommensurable with each other. (p. 127)

Even scholars who questioned the compatibility of Ricoeur’s thought with Macron’s political stance acknowledged that Macron is a “man of synthesis, who tries to hold contradictory positions together” (Aeschimann, 2017). Applauded by some for his innovative attempts to transcend traditional divides, yet criticized by others for exploiting ambiguity, Macron exhibited a paradox mindset prior to the COVID-19 crisis.

When the COVID-19 crisis surfaced, Macron, like many other world leaders, addressed the population on many occasions, while navigating the paradox, with poles that we refer to as “saving lives” versus “preserving life as usual.” In Macron’s speeches, we identified four narratives reflecting contrasting ways of navigating paradox. While Macron had been recognized for his ability to think paradoxically before the crisis, our data suggest that in the face of this unprecedented crisis, he was forced to improvise. He would first engage each pole separately before striving for a more balanced trajectory between both poles, seeming to satisfice rather than optimize as he progressively learned to embrace the inherent uncertainty of the situation (Table 1).

OPEN IN VIEWER

Table 1.

Overview of the phases.

Four narratives that emerged throughout the crisis

Turning points

While Macron was known for his ability to see opposites and bring them together in a both/and vision, our data suggest that the paradox of “saving lives” versus “preserving life as usual” in a time of crisis challenged such a vision. Indeed, the paradox navigation we described entailed brutal shifts in narratives akin to turning points, that is, “short, consequential shifts that redirect a process” which are often “abrupt” and “chaotic” (Abbott, 2010: 249, 258). These turning points represent profound changes in trajectories within a temporal sequence (Vaara et al., 2016). Specifically, they can result from sudden and unexpected incidents with potentially high societal impacts such as major crises and disasters, which force leaders to improvise (Mendonça et al., 2004). They mark the beginning and end of each phase and are characterized by strong contrasts and discontinuities that set radically different trajectories. The quiet normality of Phase 1 gave way to Phase 2, characterized by an abrupt shift to a war and crisis narrative. As the virus receded, the narrative of Phase 3 reinstated some form of normality, which again was disrupted by the return of the “second wave” in Phase 4 and the continuous blurring of boundaries between life as usual and saving lives. Turning points are also characterized by their abruptness, especially before each incoming virus wave. For instance, only a few days separated the Phase 1 narrative of normality from the discourse of emergency in Phase 2. Similarly, the second wave, even though anticipated as a possibility, emerged so quickly that it took Macron by surprise. The resulting radical shifts, taken at the last moment, is typical of improvised actions, requiring adaptation “in the moment rather than reflection beforehand” (Cunha et al., 2012: 268).

Fog of uncertainty

Our data suggest that paradox navigation was heavily hampered by a fog of uncertainty. This metaphor suggests that most actions in war “are wrapped in a fog of greater or lesser uncertainty” (Von Clausewitz, 1943: 32), making it difficult to develop situational awareness given the many unknowns (e.g. adversary capability, intentions, random events). Specifically, Clausewitz insisted on the unpredictability of war (Beyerchen, 1992). In our case, the fog of uncertainty prevented Macron from correctly predicting future trajectories, placing him in a situation where improvisation was required, because “the future cannot be ‘domesticated’ or reduced to logical continuity” (Cunha et al., 2012: 268). Indeed, predictions constantly ended up being contradicted by facts. To illustrate, the announcement before the first wave that there would not be any major sudden changes in terms of lockdown measures was contradicted by a hasty lockdown a few days later. Similarly, after the first wave, an announcement that the country would be ready for a potential second wave and that no national lockdown would take place was later contradicted by the facts. Another striking example is the evaluation of the duration of the crisis, initially estimated as lasting only a few weeks or months.

Clausewitz also insisted on the nonlinear nature of war with a “disproportionately large role of the least important of individuals and of minor, unforeseeable incidents” (Beyerchen, 1992: 77). In our case, the fog of uncertainty prevented Macron from correctly interpreting weak signals (Mendonça et al., 2004), which had exponential consequences later on. In Phase 1, as developments unfolded in Italy and the first few cases were detected in France, Macron’s stance was that life should still be conducted as usual. Similarly, the arrival of the second wave came as a surprise. In both cases, the nonlinear, exponential growth of the virus made paradoxical tensions hard to manage as previously hidden dynamics suddenly became apparent. The necessity to save lives appeared abruptly at the last moment, the same way a cliff may lurk in the fog before being spotted just before a fall. Thus, the fog of uncertainty explains why paradox navigation was characterized by abrupt turning points.

Chaotic learning

Our data suggest that Macron progressively learned to better navigate the paradox. However, rather than being built in an orderly fashion on solid ground, this learning was difficult, erratic; it shifted over time and remained prone to destabilization as the situation evolved. We thus characterize this learning as “chaotic,” thereby stressing its nonlinear nature and emphasizing the sensitivity of the entire system to small fluctuations (Van de Ven and Poole, 1995). The difficulty associated with integrating lessons from the past is manifest in repeated mistakes. In Phases 1 and 3, the situation was described as “a bit peculiar.” In both cases, these phases were followed by the rapid onset of new waves of the virus that took everyone by surprise (in Phase 2 and Phase 4). Such difficulty is typical of improvisational learning (Miner et al., 2001), which happens in the moment. Specifically, the episodic nature of the pandemic made it prone to foster unlearning between each wave, with lessons forgotten after the fact (Cunha et al., 2015).

Still, with time, Macron learned from past experience progressively. First, he repeatedly acknowledged past failures. For instance, in Phase 4, the chosen solution when facing the second wave was less extreme than in Phase 2. Macron explicitly justified his choice by invoking lessons learned from the past. He leaned more toward a both/and strategy of embracing both poles (Lewis, 2000), adopting a satisficing goal (Simon, 1957) rather than an optimizing one. Also, he began to rely much more on figures and data to justify his choices, signaling an accumulation of knowledge on the topic. Finally, he signaled a much more “humble” attitude: rather than setting up specific measures, he adopted a more contingent approach, with a narrative paving the way for multiple paths. Overall, such progress suggests a transition from improvisational to trial-and-error learning (Miner et al., 2001). However, in contrast to prior works suggesting an orderly sequence of learning (Bingham and Davis, 2012), the constantly fluctuating nature of the environment and its sensitivity to changes that may precipitate new dynamics—such as new virus strains—make any consolidation of knowledge difficult.

Finally, chaotic learning is characterized by the difficulty of engaging in vicarious learning. For example, in Phase 1, vicarious learning opportunities were missed as the case of Italy was discounted. More generally, the fact that other European countries weathered the same storm quasi-simultaneously made vicarious learning next to impossible, as illustrated in Macron’s mentions of other countries experiencing the same setbacks.

In sum, we observed a learning process that was slow and unstable, with setbacks and episodes of unlearning typical of situations with high improvisation (Cunha et al., 2015; Miner et al., 2001). Only when Macron was able to recognize the chaotic nature of the crisis and relinquish the illusion of being able to tame the virus and accurately predict the future—thereby acknowledging the fog of uncertainty—did we observe a more balanced navigation strategy; it was no longer marked by abrupt turning points, but was based on prudent, contingent fine-tuning aimed at satisficing both poles.

Contributions to paradox theory

Practical implications for management learning and improvisation

Our model also has implications for management learning and improvisation. Indeed, our work resonates with prior works showing that improvisation can lead to multiple forms of learning as it “may gradually metamorphose” (Cunha et al., 2015: 515). Chaotic learning suggests that progressing toward a more structured form of learning is possible (e.g. improvisational toward trial-and-error learning), but that this progression is fraught with difficulties and is constantly prone to setbacks, given the highly unstable and dynamic system within which it occurs (Van de Ven and Poole, 1995). As prior research suggests, “in highly interactive systems, individual improvisational actions have a greater chance of causing inconsistency” (Miner et al., 2001: 328). We believe that management learning and education thus could benefit greatly from further examining paradoxical systems, which by nature are highly interactive and even encourage leaders to consistently provide inconsistent responses (Smith, 2014).

As paradox scholars teach paradox navigation, they may highlight that in unstable systems in which paradoxes are extremely heightened, new forms of situated learnings emerge—learnings in which the here and now supersedes. Our study suggests that teaching students and managers to acquire a paradox mindset would be but a first step. Actors should not only be made aware of the risk of losing this mindset and thus be taught to continuously “refresh” their views (Lüscher and Lewis, 2008), they should also learn to ponder reflexively on their own ability to reactivate paradoxical thinking when facing new tensions in unusual, surprising situations. They should also be aware that they may need idiosyncratic knowledge and skills specific to a given context (Brown, 1994). Improvisation by nature is adapted to a specific context (Cunha et al., 2015, 2020), and in particular, improvisational skills such as adaptability and resilience are key in contexts characterized by high uncertainty and time pressure (Crossan et al., 2005). In addition, leaders need to learn to develop “real time foresight,” where planning is complemented by quasi-simultaneous immediate execution (Cunha et al., 2012: 268), for instance, by learning from both failed and successful previous improvisations (Cunha et al., 2015). In sum, our study confirms previous insights that leaders should learn to develop their peripheral vision—paradoxically looking wherever attention is lacking—and to scan the environment for weak signals, as well as to act swiftly when detecting them (Cunha et al., 2020; Mendonça et al., 2004). All in all, managers should be taught to embrace uncertainty as a complement to the traditional organizational focus on predictability and routines (Cunha et al., 2015).

Finally, we have revealed implications for improvisation by nuancing its paradoxical nature. Prior work has underlined the paradoxical quality of improvisation as it combines “order and chaos” and “planned and spur-of-the moment action” together (Tabesh and Vera, 2020: 2236). Our study builds on these insights and suggests that both preparing an intellectual structure (i.e. developing a general paradox mindset before the crisis) and responding extemporaneously (i.e. adapting freely to a specific, quickly evolving situation) go hand in hand in improvisation. Thus, paradoxically, leaders should be prepared to be unprepared and combine preparation (i.e. a paradox mindset) with embracing flexibility in the face of unpredictable events.