Narratives of Project Risk Management: From Scientific Rationality to the Discursive Nature of Identity Work

The dominant narrative of project risk management pays homage to scientific rationality while conceptualizing risk as objective fact. Yet doubts remain regarding the extent to which the advocated quantitative techniques are used in practice. An established counternarrative advocates the importance of intuition and subjective judgment. New insights are developed by conceptualizing risk as a narrative construct used for the purposes of identity work. Project-based practitioners are seen to mobilize resources from competing narratives to meet the transient expectations of those with whom they interact. Ultimately, they tend to emphasize approaches that sustain their ascribed identities as custodians of rationality.


Introduction
Risk management is routinely recognized as a key component of effective project management. It is routinely flagged as such within the codified professional bodies of knowledge promoted by associations such as the Association for Project Management (APM, 2019) and the Project Management Institute (PMI, 2017). There are also numerous textbooks on project risk management that prioritize the use of quantitative tools and techniques Edwards & Bowen, 2005;Kendrick, 2015;Raydugin, 2013). The accepted doctrine of project risk management tends to focus on core principles and critical success factors while emphasizing the importance of a systematic and disciplined approach (cf. APM, 2010;PMI, 2017). The literature is overwhelmingly prescriptive in that it is primarily concerned with what practitioners should do (Senesi et al., 2015). Yet empirical research, which seeks to demonstrate causality between the use of project risk management and project success, remains at best stubbornly inconclusive (Willumsen et al., 2019).
It is further widely accepted that the actuality of risk management invariably differs from current mainstream prescriptions (Olechowski et al., 2016;Papke-Shields et al., 2010;Taylor, 2006;Kutsch & Hall, 2009). There nevertheless remains a paucity of research that explores the reasons for this widely observed variance between theory and practice. There is even less research that focuses on the praxis of project risk management, in other words, the ways in which theory is understood and enacted in practice (cf. Thomas et al., 2012). Hence the research question that arises is as follows: Why are the traditional quantitative techniques of project risk management continuously reproduced in the prescriptive literature if they are so rarely used in practice? The guiding proposition is that practitioners utilize the available narratives of project risk management for the purposes of identity work. Our fundamental point of departure is the contention that risk is best understood as a construct that humans overlay on the world around them (Jasanoff, 1983). We build on this essential idea by conceptualizing risk as a narrative construct. On this basis we seek to make a theoretical contribution to current understanding that extends beyond the prescription of supposed best practice. The adopted narrative perspective forms part of the broader linguistic turn in the social sciences, which views social phenomena as being constituted through language (cf. Gabriel, 2004). Hence risk becomes a word that is mobilized for the purposes of sensemaking rather than a substantive thing to be managed (Taarup-Esbensen, 2019). Zhang (2011) notably distinguishes between two schools of risk analysis on epistemological grounds. The first sees risk as objective fact, whereas the second construes risk as subjective and socially constructed. Similar arguments have also been proposed based on alternative modes of thinking (e.g., Slovic et al., 2004;Kahneman, 2011). From a narrative perspective, we argue that such modes of thinking are best understood as offering alternative storylines that appeal to different constituencies. One tends to be favored by mathematicians and engineers and the other by social scientists (Jasanoff, 1983). However, such categories are neither fixed nor mutually exclusive. We contend they are better understood in terms of the identities that individuals seek to create for themselves. But we also recognize the strong abiding expectation that project managers should present themselves as objective, rational, and disciplined (Hodgson, 2002).
Of central importance to the adopted perspective is the recognition that narratives are directly implicated in the way organizations are constituted and continuously re-negotiated (Frandsen et al., 2017). Hence the lived reality of practices, such as project risk management, is experienced and enacted through the medium of narrative (cf. Rantakari & Vaara, 2017). We further see the narratives and stories mobilized by individuals as being critical components of identity work (Alvesson et al., 2008;Ibarra & Barbulescu, 2010). In this respect, the described research echoes previous studies, which emphasize the importance of identity work within the context of project-based organizations (e.g., Green & Sergeeva, 2019;Sergeeva & Winch, 2021).
This article is structured as follows. Initially, we rehearse the criticisms most often directed at the accepted doctrine of project risk management as derived from its origins in probability theory. A broader framing is then provided by addressing the supposed relationship between uncertainty and project complexity. Coverage includes the sociopolitical complexity that invariably characterizes projects with multiple participants. We provide further contextualization by introducing the emergent concept of risk governance. Thereafter, we offer an extended justification for seeking to understand project risk management from a narrative perspective. Attention is given to the multiple arenas within which project risk management is enacted and to the contention that practitioners mobilize discursive resources from narratives that are seemingly in competition. We also provide a more detailed explanation of our claim that the constituent interactions can be meaningfully understood in terms of identity work. The final section describes the empirical research. The research design is described and justified with reference to the adopted theoretical perspective. We then present three empirical narratives of project risk management as derived from a multiparticipant focus group. The article concludes with a summary discussion of the issues arising along with our conclusions in respect of the stated research question.

Probability of Occurrence
The origins of project risk management are shaped by the perennial quest for rational decision-making in conditions of uncertainty. Advocated methods, such as decision trees and risk registers, typically comprise the assignment of subjective probabilities to an identified set of envisaged events. Rational choice is further seen to be dictated by the product of the assigned subjective probability and the estimated impact of the event in question. The intellectual antecedents of this broad approach lie in Savage's (1954) concept of subjective expected utility (SEU). The overall persuasiveness of SEU among the advocates of project risk management remains remarkably intact despite sustained criticism (Winch & Maytorena, 2011). Many such criticisms notably predate the codification of project management as a discipline. For example, Tversky and Kahneman (1974) famously demonstrate how the adopted heuristics for assessing subjective probability often lead to severe and systematic errors. Their subsequent research has been equally influential in establishing how the evaluation of both probabilities and outcomes is crucially dependent upon the way decisions are framed (Tversky & Kahneman, 1981). Winch and Maytorena (2011) further suggest that the inappropriate use of project risk management techniques may be part of the problem rather than part of the solution. Others have argued that the term 'risk' has itself become an obstacle to informed decision-making (Dowie, 1999;2000). There is also a recurring tendency among practitioners primarily to focus on downside threats, despite numerous exhortations to the contrary (e.g., APM, 2010;Aven, 2017;PMI, 2017). Ward and Chapman (2003) likewise argue that accepted definitions of risk tend to focus too narrowly on the probability of occurrence of defined events or circumstances. As such, they distract attention from more systemic sources of uncertainty that cannot be decomposed into discrete singular events with defined probability distributions (Dowie, 1999;Williams, 2017). Williams (2017) notably moves beyond probability of occurrence by explicitly linking project risk to the prevailing level of systemic complexity. Numerous previous attempts to classify sources of project complexity have been developed with the aim of improved managerial responses (e.g., Baccarini, 1996;Geraldi et al., 2011;. Such attempts are influenced to a greater or lesser extent by the scientific tradition of complexity theory. Complex physical systems are commonly held to display structural complexity relating to the number of fixed elements. In contrast, complex adaptive systems are seen to comprise dynamic networks of interactions, the behavior of which is unpredictable (Holland, 2014). The uncertainty relating to complex dynamic systems is hence primarily emergent, and thereby less easy to define in terms of specific attributes (cf. Geraldi et al., 2011). Nevertheless, there is recurring confusion as to whether the uncertainty is an emergent property from within the system boundary or whether it arises from the broader environment within which the system operates. Much of course depends on where the boundary of the system is drawn and by whom.

Project Complexity as a Source of Uncertainty
Given the absence of any detached objective standpoint, the debate has more recently shifted toward the subjective interpretation of complexity (Maylor & Turner, 2017). The difficulty here is that the reality that practitioners see is forever influenced by the metaphorical lenses they use for the purposes of sensemaking (cf. Morgan, 2006;Tsoukas, 1994). Maylor and Turner (2017) further argue that the relationship between perceived complexity and managerial response is best understood as recursive. Hence practitioners' interpretations of project complexity are not only dependent upon the adopted sensemaking frameworks, but the frameworks themselves have systemic consequences. We would have much sympathy with this argument, but there is notably little emphasis on how such processes might be shaped by shifting notions of self-identity. We would further argue that terms, such as complexity, uncertainty, and risk, are ultimately best understood as narrative constructs that are mobilized by practitioners in different ways for the purposes of sensemaking (cf. Taarup-Esbensen, 2019).

Sociopolitical Complexity
Several of the above sources also refer to the sociopolitical complexity of projects that invariably arises from the involvement of multiple participants (e.g., Geraldi et al., 2011;Maylor & Taylor, 2017). In common with other supposed categories of complexity there is again a recurring problem of definition. There is also precious little guidance on how sociopolitical complexity might be better addressed, other than vague recommendations about expending more effort on managing senior stakeholders (cf. Geraldi et al., 2011). Advice of this nature too easily neglects the fraught political difficulties often involved even in agreeing on who the senior stakeholders are. Moreover, there is little guarantee that any such negotiated consensus would remain in place over time.
It can further be argued that the courses of action available to project managers are inevitably constrained by issues of political capital (Clegg & Courpasson, 2004). Similar arguments apply to the choices available to project risk managers. Managers may well be highly motivated in accordance with stated project objectives, but they often operate in matrix organizations where different objectives pull in different directions (Larson, 2004). Hence the very way 'risks' are framed can be construed as a politicized exercise dependent upon the agreement of competing interest groups.
Individual managers are also likely to act in a manner that enhances their own reputation with a view to securing future employment. In short, project team members invariably have aspirations of their own entirely disconnected from the formally stated project objectives. Such aspirations will inevitably morph in accordance with the availability of career opportunities elsewhere. They may also change in accordance with the personal circumstances of those involved and in accordance with their shifting sense of self-identity. Hence the uncertainties that arise from sociopolitical complexities differ from those associated with the scientific tradition of complexity theory (cf. Holland, 2014). As such, they are deserving of separate conceptualization. But the important point is that sociopolitical complexity ascribes project participants with the agency to act in accordance with their own interests.

Risk Governance
Given the extensive criticism directed at the constituent techniques of project risk management, it is pertinent to ask what their advocates are purportedly striving to achieve. The answer is provided at least in part by the emerging concept of risk governance (Stein & Wiedemann, 2016;van Asselt & Renn, 2011). The concept of risk governance is concerned with the need for organizations to deal with risk responsibly. There is an acceptance from the outset that risks often cannot be calculated solely as a function of probability and effect. Hence there is an expectation that risks are multicausal and routinely surrounded by uncertainty and ambiguity (Renn, 2008). Decisions about risk are further seen to involve a complex interplay among multiple actors rather than being amenable to technocratic solution. The focus of the risk governance community therefore extends beyond prescriptive techniques toward broader institutional arrangements and associated political cultures ( van Asselt & Renn, 2011). We would argue that risk governance is especially relevant in the context of megaprojects where there is invariably a need to be accountable to multiple governance structures (cf. Gil et al., 2017).
The emphasis of risk governance on the need to manage risk responsibly accentuates the need for communication, transparency, and accountability. Boholm et al. (2012) are typical in emphasizing the contextual embeddedness of risk governance, in other words, that it takes place in contexts that are historically, spatially, and institutionally situated. The situated nature of risk governance thereby further discredits the traditional narrative that risk is subject to scientific analysis in accordance with generic principles. Boholm et al. (2012) are also typical in emphasizing the importance of experiential learning. They further see risk governance not as a set of techniques or as a framework, but something that is learned in the context of practice (cf. Nicolini et al., 2003). It is on this basis that we advocate the conceptualization of risk as a narrative construct.
Understanding Project Risk Management from a Narrative Perspective

Project Organizing Across Discursive Arenas
Advocates of the narrative turn in organization studies contend that the world is constituted by shared language and can hence only be understood through the medium of narrative (Vaara et al., 2016). The process of organizing can further be understood as being continuously enacted across a series of socially constructed discursive arenas (Weick, 1995;Currie & Brown, 2003). It follows that human actors experience the world in terms of the narratives with which they interact during their day-to-day practices (Polkinghorne, 1988). The narratives of project risk management can thereby be seen to contribute discursive resources, which help managers make sense of the complex realities within which they operate. Such resources include persuasive sound bites derived from narratives such as A Guide to the Project Management Body of Knowledge (PMBOK ® Guide) -Sixth Edition (PMI, 2017). They also include commonly advocated tools and techniques (i.e., knowledge artifacts) such as decision trees and cause-and-effect diagrams. It is further widely recognized that project uncertainty does not necessarily diminish over time (Jaafari, 2001). Hence project managers continuously seek to make sense of the factors that impact upon project success throughout the project life cycle and continuously adjust their strategies accordingly. The conceptualization of project risk management as a process of discursive contestation enacted across multiple organizational arenas is especially pertinent in the context of megaprojects. Such projects are routinely characterized by multiple governance structures populated by changing coalitions of stakeholders (van Merrewijk et al., 2008;Flyvbjerg et al., 2003). In such circumstances, the notion that risks can be identified in advance and probabilities assigned based on a scientific algorithm would seem at best naive. What becomes more important is the projection of a narrative of project risk management that provides an overall sense of direction to a multitude of different parties. Yet the narrative must also be sufficiently malleable to be able to respond to the shifting concerns of different stakeholder groups. It follows that effective project risk management requires a leadership narrative that continuously unfolds across extended time lines (cf. Sergeeva & Winch, 2021). The existence of such a strong legitimizing narrative of project risk management hence becomes an essential governance requirement.

Discursive Resources
For the purposes of understanding practice, it is useful to focus on narratives as providing practitioners with alternative sets of discursive resources. Those who seek legitimacy for the selfascribed role of technical expert would draw resources from the narrative of risk as objective fact. Alternatively, those who experience project risk management as an inherently politicized process would be more likely to mobilize discursive resources from the narrative of risk as a social construct. Hence the practice of project risk management becomes inseparable from the adopted narrative. For example, the adoption of the language of social constructivism as embedded within so-called soft approaches creates an expectation of a more consultative style with an orientation toward consensus building (cf. Drummond, 1996). Practitioners of course would make pragmatic use of multiple sets of resources in accordance with the expectations of those with whom they interact at different points in time. In many cases, they would seek to establish their initial credibility by drawing from the rational narrative of risk as objective fact. However, practitioners thereafter would need to mobilize alternative narratives for the purposes of making meaningful progress. Narratives can further be seen as the medium through which managerial practices, such as project risk management, are shared within and across organizational boundaries.

Project Risk Management as Identity Work
The narrative of the iron triangle has long since been recognized as being influential in shaping what project managers prioritize (Atkinson, 1999). This provides a good example of a performative narrative that 'brings theory into being' (Gond et al., 2016). Such narratives are also recognized as performative in shaping project managers' sense of self-identity (Hodgson, 2002). Project managers routinely include phrases, such as managing risk, in their leadership narratives for the purpose of instilling a shared sense of direction across diverse project teams (cf. Havermans et al., 2015;Sergeeva & Winch, 2021). This is equally true of the more anecdotal stories they tell for the purposes of collective sensemaking (Weick, 1995). We would argue that many project managers take on the mantle of risk management as an essential part of their self-identity. This is routinely encouraged by the prominence given to risk management in the codified 'bodies of knowledge' (e.g., APM, 2019; PMI, 2017). Some practitioners may provide an in-house risk management service, perhaps working across multiple projects within a centrally managed support function. Others may transition over time from the role of generic project manager to that of specialist project risk manager; a few may even transition in the opposite direction. However, it is reasonable to suppose that practitioners always strive for a degree of alignment between the roles they perform and the self-identities that they create for themselves.
Those tasked with responsibility for risk governance would undoubtedly create different self-identities from those with responsibility for project delivery. Even more crucially, the collective expectation would be continuously shaped by the different audiences with whom they interact. In the absence of other overriding sources of legitimacy, the established narratives of project risk management would remain important points of reference for all involved. The normative narratives promoted in the prescriptive literature are directly implicated in shaping what gets talked about and what gets ignored. They are also implicated in the techniques (i.e., knowledge artifacts) that are mobilized for the purposes of analysis. This remains true irrespective of the extent to which such analysis tends toward the symbolic (cf. Drummond, 1996).
It is further important to recognize that identity work is invariably fluid and temporal. Identities are rarely fixed and immutable but tend to be continuously (re)negotiated through interaction with others. Hence the issue of self-identity becomes of central importance to how project risk management is enacted. Ultimately, it is identity work of this nature that influences how people make sense of the practice worlds within which they operate and the actions they prioritize (Alvesson et al., 2008). This applies equally to the generic professional discipline of project management as it does to specialisms such as project risk management (cf. . It also applies to the temporal transitions that characterize project organizing (Linehan & Kavanagh, 2006). Hence it becomes important to understand the extent to which mainstream narratives of project risk management are reflected in the stories mobilized by those involved.
Our contention is that the quantitative version of project risk management is best understood as a performative narrative that is mobilized by its advocates for the purposes of bringing project risk management 'into being' (cf. Gond et al., 2016). In contrast, the so-called soft approach comprises a less formalized counternarrative that struggles for equivalent recognition. One of the difficulties is that there remains little guidance on how soft project risk management might meaningfully be enacted.

Problem-Structuring Methods
In seeking a persuasive narrative of soft project risk management, it is initially pertinent to look at the literature on problemstructuring methods (PSMs) (e.g., Mingers & Rosenhead, 2004;Rosenhead & Mingers, 2001;Pidd, 1996). PSMs are considered relevant because they have specifically arisen from a sustained critique of the quantitative hard paradigm of management science. They have further been specifically developed to provide guidance in situations characterized by high levels of uncertainty (cf. Rosenhead & Mingers, 2001). PSMs are also of note in placing considerable emphasis on the importance of organizational and individual learning (Pidd, 1996). In the specific context of projects, Winter (2006) argues in support of a greater use of such methods, especially during the early stages of projects where objectives are so often unclear and different constituencies have conflicting aims. The strategic choice approach (SCA) is rendered especially relevant by its explicit strategic focus on the management of uncertainty (cf. Friend & Hickling, 2004;first published 1997). It is also notable for its underpinning recognition of the interconnectedness of decision problems. It would hence seem to resonate with much of the preceding discussion regarding the limitations of existing conceptualizations of project risk management (cf. Williams, 2017). The approach has previously been advocated for use during the front end of projects by Green (2001) and is alluded to in passing by Ward and Chapman (2003). However, with these two notable exceptions, it has to date attracted little attention from within the projects community.

Bringing Precision to the Signifier 'Soft'
SCA is perhaps best understood as a 'soft' version of decision analysis in that it explicitly recognizes that complexity often results from multiple stakeholders with different viewpoints. (Mingers, 2011). However, it is important to bring a greater degree of precision to the terminology of 'soft' versus 'hard' than often prevails within the project management literature (cf. APM, 2018; Thomas et al., 2012). We follow the view that traditional methods of decision analysis are usefully classified as 'hard' on the basis that the advocated models are held to be representative of an assumed objective reality (cf. Pidd, 1996). In contrast, the models produced within the tradition of 'soft' decision analysis are primarily seen as tools for thinking. The overriding aim of the latter is to facilitate reflective learning and collective sensemaking (Rosenhead & Mingers, 2001). From this perspective, a greater emphasis is given to the modeling process (Phillips, 1984). Such an interpretation aligns directly with Bredillet's (2010) integrative approach to project management, which sees modeling primarily as a means of understanding. However, it is important to emphasize that we see the distinction between 'hard' and 'soft' primarily as a means of understanding different forms of narrative-and the arguments embedded there within. 'Hard' narratives would hence be characterized by the terminology of positivism, whereas 'soft' narratives would place more reliance on the language of social constructivism. In common with Thomas et al. (2012), we remain cautious of the extent to which these descriptors can be extended to the classification of practice.

Managing Uncertainty
The published narratives of SCA are notable for emphasizing the participative and iterative nature of decision-making as structured around four modes: • Shaping the decisions to be addressed; • Designing alternative courses of action; • Comparing the identified courses of action; and • Choosing an agreed course of action (Friend, 2001).
These specified modes of decision-making can be loosely equated with spatiotemporal arenas, whereby different participants come together for the purpose of meaningful action. Of key importance is the framing (and continuous reframing) of decisions in terms of their perceived urgency and importance. Of further note is the accompanying conceptualization of three different types of uncertainty relating to (1) the broader operating environment, (2) guiding values, and (3) related decisions areas. We would further hold it to be important that the process is enacted in the situated language of the participants, rather than being obscured by the quantitative representations associated with project risk management. Participants would include those with direct responsibility for project delivery who would be expected to take ownership of the agreed outcomes. Decisions considered urgent would on occasion have to be made despite high levels of prevailing uncertainty. But, more generally, the agreed outcomes would comprise strategies for managing uncertainty over time (Friend & Hickling, 2004). SCA would hence seem in principle to offer a guiding narrative that offsets many of the recurring criticisms directed at quantitative approaches to project risk management. Yet, with few exceptions, it remains largely ignored within the project management community, despite its obvious relevance to the management of uncertainty. What it seemingly lacks is the same level of political support enjoyed by more traditional approaches to project risk management.

Research Design
Methodology The adopted theoretical perspective contends that narratives are constitutive of reality rather than merely representative (Bruner, 2002;Weick, 1995). It would hence be epistemologically inconsistent to adopt a research method aimed at exploring an assumed external objective reality. Of further importance is the underlying recognition that project organization is an activity in a perennial state of continuous becoming (Musca et al., 2014;Linehan & Kavanagh, 2006;Sergi et al., 2020). Hence the meaning of project risk management becomes subject to continuous renegotiation, likewise the roles and self-identities of those involved. Previous studies involving narrative methods tend to rely on semistructured interviews (e.g., Havermans et al., 2015;Sergeeva & Zanello, 2018). However, despite their continued popularity, such interviews cannot be construed as providing a neutral account of how the interviewees interpret the subjective world within which they operate (Hopf, 2004). They are more realistically seen as arenas where both interviewee and interviewer engage in identity work, with inevitable consequences for the course and content of the interaction (Cassel, 2005). Hence Holloway and Jefferson (2008) promote the specific genre of a narrative interview, whereby both parties are seen to be actively involved in the co-production of empirical data. Cicmil et al. (2006) notably highlight critical dialogue between researchers and practitioners as an essential means of understanding the actuality of projects.
It follows that narrative interviews differ from standard interviews not only in terms of their purpose but also in terms of underlying epistemology. It is well recognized that the appropriate choice of research method is irrevocably shaped by the adopted theoretical perspective (Van Maanen et al., 2007). Such theoretical shaping is readily evident in Veenswijk and Berendese's (2008) study of how new working practices are constructed through project narratives. They notably rely on interviews while explicitly recognizing the importance of multiple voices (i.e., plurivocality) and the processes through which different stories are socially constructed over time. Musca et al. (2014) likewise focus on the co-construction of project narratives, but through the lens of a longitudinal ethnographic study. In contrast, Boddy and Paton (2004) rely on secondary case studies for the purposes of illustrating competing project narratives. In short, there is an existing trajectory of narrative research methods within the context of project studies.

Focus Groups as a Research Method
The concept of a narrative interview can be extended to include multiple participants. Focus groups of this nature have long since been utilized by qualitative researchers (Morgan, 1996). They are also a commonly adopted research method within the context of project studies (e.g., Fisher, 2011;Yu & Leung, 2015). Examples include their specific use by Hodgson et al. (2011) for the purpose of observing the social construction of work identity. Other precedents for the use of focus groups as a means of accessing narrative accounts include Hampton's (2004) study of public participation in policy making. The use of focus groups can be seen to be broadly consistent with the principles of narrative interviews while at the same time offering important additional advantages (Ellingson & Ellis, 2008). Interviews are often limited to a binary interaction between respondent and researcher. In contrast, focus groups require participants to justify themselves in front of an informed audience, which crucially includes their peers. In this respect, focus groups approximate toward the multiparticipant discursive arenas within which project risk management is so often enacted. Given the focus of interest on identity work, it was considered especially important that the respondents should interact with a diversity of peers from within their own practice worlds (cf. Nicolini et al., 2003).

Focus Group Design
The focus groups were conducted within the context of a participative one-day workshop at the headquarters of the Turkish Contractors' Association (TCA) in Ankara. The event was promoted through the auspices of the TCA as a knowledge-sharing workshop. Nineteen participants attended from a diversity of firms within the Turkish construction sector. Turkish firms were targeted because of their recognized capability for operating in high-risk markets (cf. Duman et al., 2019;Öz, 2001). Four of the companies represented were at the time named in the Top 250 list of global contractors maintained by the Engineering News Record (ENR).
The workshop was structured in three distinct phases, with adequate breaks for lunch and refreshments. Each phase culminated in a facilitated focus group-or multiple focus groups conducted in parallel-used for the purposes of data collection. The adopted agenda was specifically designed to elicit co-produced narratives in accordance with the targeted research question. Proceedings commenced with a presentation from the research team comprising a critique of the traditional techniques of 'hard' project risk management. Particular attention was given to the distinction between the techniques advocated in the literature and those more commonly used in practice. Participants were thereafter invited to describe the approach to project risk management that prevailed within their companies. Methodologically, it is important to emphasize that there was no attempt to produce any single monolithic account; a multiplicity of contributions was positively encouraged in accordance with the principle of plurivocality.
After the initial focus group discussion, the event was given new momentum by a presentation from the research team outlining the alternative 'soft' narrative of project risk management as exemplified by SCA. This was positioned in opposition to the previously shared 'hard' narrative with a corresponding emphasis on the language of uncertainty, rather than the more restrictive terminology of risk as critiqued by authors such as Ward and Chapman (2003). Particular emphasis was given to the importance of representing perceived uncertainties in the favored terminologies of the participants. Thereafter, a participative role play exercise was conducted using case study data derived from the researchers' consultancy experience. The 19 participants were subdivided into four subgroups, each of which was asked to simulate the initial stages of a project risk management workshop using the shaping techniques derived from SCA. On this basis, each group sought to identify sources of uncertainty relating to different aspects of perceived project complexity. They were also asked to derive strategies for the management of uncertainty over time, and thereafter to communicate their findings in a plenary session with the use of visual representations. The resulting discussion realized further data in narrative form, which lent itself to analysis.
The event was completed by a final presentation by the research team emphasizing the limitations of soft project risk management, especially with respect to the dangers of prioritizing a negotiated consensus over the importance of facts and logic. The previously articulated limitations of hard project risk management were also briefly revisited. It was emphasized that there is no generic recipe for best practice and that rigor comes from continuously questioning taken-for-granted assumptions. The challenge was presented in terms of finding new ways of thinking beyond the constraints of instrumental rationality. Brief coverage was also given to the emerging concept of risk governance and the need to manage risk responsibly. It was further suggested that different audiences are likely to be persuaded by different narratives, and that any such acceptance would be at best temporal. These provocations prompted additional discussion thereby providing additional opportunities for data collection.
We would further emphasize that each participant was consistently held to account based on the plausibility of their arguments. This task fell in part to the members of the research team, but it was also a task willingly performed by the participating practitioners. In this respect, the focus groups sought to replicate the way practitioners are routinely held to account in their day-to-day practice worlds.

Participants
Participants included representatives of senior management in addition to those with specific responsibilities in project risk management. Typical job titles included: Risk Manager, Vice President, Deputy General Manager, Quality Manager, Contracts Manager, Business Manager, Technical Manager, and Construction Coordinator. Participants tended to be highly educated and the majority had university degrees from Turkey's top engineering schools. Several had also gained postgraduate qualifications from leading universities in the United States. The event sought to explore the extent to which the accepted narratives of project risk management are reflected in the anecdotal stories mobilized by the participants. The event was conducted primarily in English, although some of the participants on occasion chose to communicate in Turkish. It was jointly facilitated by the two named authors, and the discussions were recorded and fully transcribed. The conversations conducted in Turkish were subsequently translated into English. It is worth mentioning that the two academic leads both have considerable consultancy experience in project risk management, thereby lending them additional credibility in the eyes of the industry participants.

Mode of Analysis
The adopted mode of analysis comprised a combination of thematic and structural approaches Riessman, 2008). The initial thematic analysis commenced with a process of familiarization comprising multiple readings of the narrative transcripts. This was followed by a search for recurring patterns, subsequently developed into tentative themes. The emergent themes were continuously tested against the original data and revised accordingly (cf. Polkinghorne, 1988). We were careful throughout to remain open to unanticipated findings and to revise preconceptions as necessary (cf. Reichertz, 2004).
The subsequent structural elements of the analysis differed in that they focused on the adopted plot structures (cf. Czarniawska, 2004;Riessman, 2008). Of particular interest was the way in which the recorded narratives created subject positions for the narrators themselves and for others deemed to have a significant interest. The structural analysis notably sought to access the views of the participants in terms of the roles ascribed to various parties, but it also sought additional insights by contextualizing them within a broader contextual setting . The adopted twin mode of narrative analysis hence imbued the collected data with a richness beyond that which could be achieved through more narrowly construed thematic approaches focused solely on content. The elicited data comprised the narratives used by the participants to justify their adopted approaches to project risk management in front of an informed audience of peers. This was a deliberate methodological choice in alignment with the stated research question. We could conceivably have also mobilized feedback data from the numerous project risk management workshops with which we have been previously involved. However, less targeted data of this nature would not have contributed in any meaningful way to the research question. Such data would also lack plausibility once separated from the embedded contextual settings within which it was produced. It must further be recognized that we were ourselves inevitably implicated in the co-construction of the findings. It would hence be misrepresentative to present the analyses as if they were entirely independent of any subjective interpretation. Such a position becomes inevitable once narratives are accepted as constitutive of an ever-evolving co-constructed reality (cf. Tsoukas & Chia, 2002).

Empirical Findings Preamble
As described in the previous section, the focus group discussions were initially framed by a short presentation on the traditional techniques of hard project risk management. Participants were thereafter invited to describe their own approaches to risk management. As might be expected, some were more vocal than others. It is important to emphasize that there was no expectation that the shared narratives were in any way representative of practice or uninfluenced by the presence of the researchers. They can however be taken to provide meaningful insights into how the narrators might routinely justify their approach to others and the self-identities they create for themselves. The participants did on numerous occasions notably draw from the formalized narratives enshrined in the preceding literature review. For the sake of brevity, the lead voices recorded as follows belonged to a relatively small number of participants. However, it is important to recognize that the narratives were co-constructed by all those present, not least because any points of disagreement were so vigorously contested. The narratives hence comprise multiple data points and as such are much richer and more nuanced than singleperson interviews. They also notably morphed over time and should not therefore be judged on the basis of internal consistency.

Imposing Discipline
Yusuf was the most vocal participant in the group and the first to respond to the invitation to contribute. He was a specialist risk manager with over 21 years of experience working on construction projects. Yusuf started by describing how his company had made the decision to implement a form of enterprise risk management. This had seemingly involved the establishment of a dedicated risk group for every project. Such groups were apparently routinely tasked with the identification, monitoring, and analysis of perceived risks. Emphasis was given to the completeness of coverage, with an implied subtext that nothing was left to chance. He was further careful to describe risk management as a regular activity performed on every project, invariably commencing at the tendering stage. He described how initially estimators are involved together with several experienced project managers. Yusuf went on to describe how such sessions often include representatives from functional units within head office such as legal, contract, and technical. He explained how there was a particular emphasis on contractual risk and risks associated with unforeseen ground conditions. Following contract award, workshops were said to be conducted on a regular basis at least every two months. There was a particular focus on the importance of maintaining the risk register. It was claimed that probabilities were regularly reassessed in terms of their impacts on time and cost. Reports are apparently sent to top management with the assessed impact of the most critical risks being flagged for their attention. The emphasis was consistently placed on regular activity leading toward regular reporting. However, it was striking how the responsibility of the risk group apparently ended with the task of reporting. The responsibility for taking subsequent action seemingly lay with the company's executive management.
Working Toward a More Enlightened Approach Despite Yusuf's initially dogmatic tone, he notably shifted the emphasis of his narrative once he felt he had established his credentials within the group as a risk expert. The second phase of the storyline emphasized how his company was attempting to implement a more enlightened version of risk management, with more emphasis given to opportunities. This secondary, more conciliatory, narrative gained notable support from the other participants. This was especially true of the distinction now made between how project risk management had been traditionally performed and the approach they were now working toward. Participants noticeably felt comfortable with a narrative that was framed in terms of a desired shift toward how they wanted to perform risk management in the future. Yusuf notably made direct reference to the latest version of ISO 9001 when describing the ambition to embed risk-based thinking across the organization. Returning to an earlier theme, he explained how each department within his company had been specifically tasked with adopting risk-based thinking. The described aim was to encourage more participation from operational personnel rather than relying solely on dedicated project risk groups.
It was notable that neither Yusuf nor any of the other participants placed any initial emphasis on the use of specific techniques. It was only in response to a direct question that Yusuf subsequently referred to the use of risk dashboards. He also seemingly took particular pride in the development of in-house risk management software, which apparently makes use of a centralized risk library. But this was accompanied with the immediate caveat that there is also a strong emphasis on experience. At this point, Yusuf seemingly felt himself caught between two narratives, with an obvious concern about the need to appear to be consistent. The centralized risk library was on the one hand heralded as a means of bypassing the need to brainstorm possible risks from scratch, thereby speeding the process up. On the other hand, however, there was an awareness that there is danger in sacrificing broader participation for the sake of efficiency.
It was further emphasized that there had recently been an attempt within the company to incorporate mitigation strategies into the recommended process of project risk management. The stated aim was to achieve a greater emphasis on alleviating the consequences of identified risks. Particular emphasis was placed on the ongoing evaluation of the effectiveness of the initiated actions with a view to revising the recorded risk scores, defined as the product of probability and impact. There was therefore an explicit recognition that project risk management needed to be better integrated into line management. Overall, it was notable that there was a consistent (and perhaps deliberate) blurring between how risk management was currently implemented and how the respondent thought it should be implemented. This same tendency was reflected by several others within the group.

Educating the Project Managers
As Yusuf's confidence grew, he progressed to a topic about which he seemingly had strong feelings-the importance of distinguishing between risks and issues. He expressed a degree of frustration that others within his organization consistently confuse these terms; for Yusuf, this was indicative of an underlying lack of rigor. The narrative here drew heavily from the traditional quantitative model of project risk management. Risks are seemingly only risks if they can be assigned a probability of occurrence, otherwise they are classified more loosely as management issues. Of further interest is the way in which Yusuf cast himself into the role of educator. He seemingly saw it as his role to ensure that project risk management is applied more diligently. He was especially keen that project managers should engage in the process more seriously and take ownership of the outcomes: "We shouldn't only work with technical managers and HSE managers. The project manager's contribution is very important because in the end they will decide what to do. " Yusuf further emphasized the importance of increasing the number of workshops, which he saw to be dependent upon an increase in the available budget. In common with most departmental heads, much of his time was seemingly spent arguing in favor of additional resources. He repeatedly cited the support of top management in his quest to encourage more workshops. The underlying inference was that it is the project managers who need to be persuaded. Risk workshops were typically described as lasting between one and two days, with Yusuf's own expressed preference being for two days. It was again suggested that project managers need to be more supportive: "Project managers should encourage us to extend the duration. Key parties should leave their job and join us to just focus on the risk register." The emphasis notably lay on the risk register and the need for wider participation by those with operational responsibility, including subcontractors.

Narrative Two
Developing Proper Tools In response to Yusuf's account, several other participants notably became more vocal. Elif was next to accept the invitation to present her company's approach. She was employed as Risk Management Director for one of Turkey's leading construction firms and she was careful to emphasize her 20 years' experience of project management in the sector. From the outset, Elif described herself as having adopted a similar approach to that adopted by Yusuf. There were however important points of difference. Elif was careful to emphasize that risk management was seen to commence with a go/no-go decision based on an evaluation of both risks and opportunities. Such an approach is apparently specified in her company's strategic plan under the label of business control. There was again an emphasis on the tendering stage, especially in terms of providing information to top management so they can decide on whether to submit a bid. There was a similar focus on participation by multiple departments, rather than allowing risk management to be the preserve of specialists. There was also an emphasis on the importance of giving equal attention to opportunities. Elif further described how risk registers are consolidated monthly by top management. The implication was that project risk management operates separately from day-to-day decision-making. Perhaps motivated by a desire not to be outdone by Yusuf, Elif also emphasized that her company had developed its own bespoke software for the purposes of recording lessons learned.
Using the Tools That are Already Available Elif further reported that risk workshops typically commence by searching through previous lessons learned. The primary emphasis therefore lay on the importance of using an analytical approach. At the same time, there was also an explicit reference to relying on the perceptions and sensemaking capabilities of project managers. In this sense, the adopted narrative sought to combine discursive resources from the two traditions of risk management. But the emphasis soon shifted back to the use of the Primavera ® risk tool, which was seen to provide a purity of analysis that more intuitive approaches could not match. This was seen to be especially useful for analyzing the knock-on consequences of possible delays. The emphasis here lay on the quantitative modeling of different envisaged scenarios, with much talk of the probability range around activity duration. There was even passing mention of Monte Carlo analysis. It seemed therefore that the focus on mitigating actions related primarily to issues of scheduling during the construction phrase rather than the risks that prevail during the front-end of projects. The view was further articulated that the engineering construction sector is often lacking in the use of such generic tools, which was viewed as something that needed to be corrected.
The overriding message from Elif-and one readily endorsed by other participants-was that the engineering construction sector does not yet quite do risk management 'properly,' hence the focus on making better use of proprietary software tools. There was also a recurring emphasis on capturing lessons learned as a means of insuring the company against the loss of key personnel. The narrative here concurs with the academic literature on knowledge management, especially the subtheme that relates to the importance of tacit knowledge.

Narrative Fragments
At this point, there was a much broader willingness to contribute, with several contributions from participants who had not yet contributed, but these secondary contributions were more in the style of stand-alone comments than coherent narratives. Yusuf seemed to enjoy an acknowledged status within the group, such that few were willing to challenge his views directly. Several were careful to stress that they had only relatively recently formally adopted project risk management, and hence saw themselves at the beginning of what they described as a journey. Mehmet, as Vice Chairman of one of the smaller represented contractors, was brave enough to offer a different perspective. He described how their projects typically have a maximum duration of one year. This was offered as an explanation for why they do not implement a formalized approach to project risk management. He was also quick to emphasize that his company worked across very different markets, characterized by very different risks: "We work in a range of very different countries. We can be working in Sub-Saharan Africa for example and at the same time working in the Middle East. Then we work in Central Asia. Completely different countries, completely different risks. " The inference was that different contexts demand different approaches, with little possibility of standardization. This seemed directly to contradict the dominant narrative toward seeking ever greater standardization. It is interesting that Mehmet did not see himself as a risk specialist, hence he did not have a vested interest in the application of risk management techniques. The representatives from the larger companies seemed broadly appreciative of his point of view, including both Yusuf and Elif. Given the overall supportive atmosphere, others started to share similar opinions while at the same time sticking with the traditional quantitative script. For example, the subsequent multiparticipant discussion stressed the importance of distinguishing between risks and known facts. Facts, it was contended, could be safely left to the project managers, but risks needed to be addressed by specialist risk managers. The argument seemed to imply that the two categories could be in some way related to an external objective world. It was further seemingly taken for granted that they could meaningfully be extracted from the context within which they are embedded.
It was at this point that Yusuf seemingly felt it necessary to re-assert his authority by reverting to a more dogmatic position. He argued that if probabilities can be assigned, then the risk must be entered into a risk register. But if probabilities cannot be assigned, it was contended that the issue of concern does not qualify as a risk. The implication was that such issues were part of the day-to-day responsibilities of project managers. Although this argument was not openly challenged, it notably failed to draw any explicit support. There was seemingly an unspoken feeling that to challenge Yusuf on this point would have been to challenge his sense of self-identity.

Soft Project Risk Management
With a view to injecting new momentum into the discussion, it was at this stage that the research team introduced the participants to the ideas of sociopolitical complexity and to the oftensystemic nature of risk and uncertainty. Emphasis was given to the idea that managerial actions often have recursive consequences, and that the framing of the issues of concern is inherently dependent upon the adopted terminology. SCA was introduced as a meaningful way of framing such interventions under the label of soft project risk management. However, it was explained that the approach is structured around the strategic management of uncertainty rather than risk as traditionally understood. Importantly, SCA was not advocated on a prescriptive basis, but more as a collective means of sensemaking based on the expressed concerns of the participants. As such, the advocated approach presented a direct challenge to the more traditional quantitative interpretations of project risk management. It also ran contrary to many of the views previously expressed by the participants. As previously described, the initial outline presentation was followed by a participative role-play exercise using case study data. Engagement in the exercise was positive and generated much debate and discussion among the participants. Following completion of the exercise, the participants were asked if their views on risk management had changed and if they would thereafter seek to enact risk management any differently. The rationale behind this question was to ascertain the extent to which the exercise had challenged their sense of self-identity as risk managers.

Natural Language
Several participants expressed the view that they could see themselves using the soft approach for risk management in the future. However, it was also clear that they primarily viewed the approach as (yet) another technique (i.e., knowledge artifact) rather than as any sort of deep-rooted challenge to the way in which they conceptualized project risk management. Certainly, there was no sudden shift of allegiance from risk management to the management of uncertainty. There was however a broad acceptance of the argument in support of allowing participants to express their concerns in their own natural language, rather than insisting that they adopt the specialist terminology of risk management. They were also accepting of the benefits of structuring concerns into prioritized clusters without any insistence on the assignment of probabilities. As had been the case previously, the feedback tended to be dominated by a limited number of individuals. On this occasion, Ahmet, Vice-President of one of the Top 250 contractors listed by ENR, was one of the first to contribute: "In my opinion, it was a very efficient process because we all had the chance of presenting our own ideas by checking some very brief information about the project. So with regards to risk, I think one of the biggest mistakes we make is that we tend to rely on certain individuals to make this analysis, rather than getting a broader perspective within the company." Ahmet clearly derived additional authority from having received a master's degree from the University of California, Berkeley. His comment implies a potential downgrading of the role of risk specialists, an implication presumably not lost on those present. Yusuf had previously been very strident in outlining his views, and it is interesting to note that his tone was now much more conciliatory: "..in our company, our technique is more or less on the hard part [of risk management]. We're trying to encourage as much as possible the wider participation of related parties and stakeholders, but not yet in the style of soft risk management." Yusuf further conceded the need for change by reference to the narrative of soft risk management: "Where we fall short in terms of soft risk management is that we are not allowing everyone to think more broadly. Most probably, we should let them think freely. And we need to ensure wider participation in addressing the full scope of everything, issues, difficulties…. everything." The above quotation represents a significant shift from Yusef's opening narrative about imposing discipline but is entirely consistent with the subsequent espoused quest to move his organization toward a more enlightened approach. He notably still sees himself in the role of allowing others to do things rather than relinquishing control.
Others also notably adopted the terminology of soft risk management for the purposes of shaping their response. The comment from Elif was typical of others: "To be interacting at the same table with a broad cross-section of employees from the company, I think it is more important to use the terminology of soft risk management. It's also considering the issues and concerns in their mind." This narrative segment is indicative of a clear acceptance of risk as a negotiated construct and the need to incorporate such representations into the process of analysis. It is important however that this view was aired fleetingly within the context of an artificially created environment. It is not therefore held to be indicative of any sort of epistemological conversion, but more as an example of the discursive interactions that characterize the day-to-day practices of risk management.

Understanding Narratives as Empirical Data
Prior to discussing the empirical issues that arise from the described focus groups, it is appropriate initially to revisit the stated aim of the research. The research question posed was: Why are the traditional quantitative techniques of project risk management continuously reproduced in the prescriptive literature if they are so rarely used in practice? It was further argued that, despite the applied nature of the discipline, there is surprisingly relatively little research into the praxis through which the accepted doctrine is understood and enacted by expert practitioners (cf. Thomas et al., 2012). Hence, the practitioner narratives outlined earlier comprise empirical data deserving of analysis.

Imposing Discipline
The described focus groups can be construed as approximating toward the discursive arenas within which project risk management practitioners routinely operate. The findings illustrate the tendency of practitioners to mobilize resources from the established narratives of project risk management. The dominant default tendency was to rely primarily on the narrative of risk as objective fact. This was especially noticeable in the opening contributions by Yusuf, with an accompanying emphasis on the supposed cognitive limitations of others. Many of the most pointed criticisms were notably directed at project managers in terms of their supposed inability to adopt the right language. It would seem therefore that there is some resistance among project managers to the narrative of risk as objective fact. Yusuf further saw his own role, and arguably even his selfidentity, to be primarily concerned with imposing discipline on others (cf. Hodgson, 2002). At times this was alternatively construed as educating them in the right way of thinking. The overriding concern was that the participating project managers should adhere to the required prescriptive protocol. That they themselves should find the process to be in any way useful was seemingly at best a secondary consideration. It was especially telling that it was considered necessary to specifically emphasize the importance of the project manager's contribution alongside that of the various technical specialists.

Playing to Different Constituencies
Although few of the respondents talked explicitly about risk governance, there were frequent references to the need to provide cogent information to corporate head office. For example, Elif specifically referred to risk management commencing with a go/no-go decision based on an evaluation of risks and opportunities. Hence she saw her role as primarily focused on the provision of information. More broadly, there seemed to be a recurring tension between serving the needs of project delivery on the one hand, and the requirements of corporate governance on the other. Such tensions would seem to be central to the lived experiences of the participants. Most of the risk managers taking part in the workshop were notably embedded within a centralized head office function. Hence they are ultimately dependent upon the patronage of those responsible for corporate governance rather than those directly responsible for site-based project management. The need to manage the expectations of these two constituencies arguably accounts for the observed tendency of the respondents continuously to morph between different narratives. Yet it must also be acknowledged that these two constituencies often in themselves comprise multiple interest groups. Even if consideration is limited to those involved in on-site activities, the consumers of project risk management invariably include a range of technical specialists. Of particular importance are those with responsibility for health and safety.

From Modes of Thinking to Modes of Narrating
Elif was notably keen to emphasize the importance of upside opportunities in addition to the more usual downside risks (cf. . Such arguments are widely wellrehearsed in the literature (e.g., APM, 2010;Aven, 2017;PMI, 2017). However, it would seem that Elif chose to emphasize the importance of upside opportunities for the very reason they were not specifically mentioned by Yusuf. Hence it was an obvious argument to make for the purposes of not allowing Yusuf to claim the space of working toward an enlightened approach entirely for himself. Elif and Yusuf can be seen therefore to have been engaged in a degree of competition, without ever quite feeling the need to publicly disagree with each other. Elif clearly gained broad approval from the audience in response to her storyline about the importance of capturing lessons learned. Such exchanges are suggestive of a political interpretation of risk management (cf. Drummond, 1996). They are also indicative of the concept of identity work in action (cf. Alvesson et al., 2008). The adopted focus on narrative further emphasizes the way practitioners continually develop ante-narratives as a means of making sense of their own experiences. Of particular note is the way such processes are carried out in interactions with others. Many of the respondents in the workshop notably shifted their position in response to the reactions of others. It hence becomes abundantly clear that practitioners do not commit themselves to any highfaluting methodological position, which they then adhere to with absolute consistency.
They were clearly more intuitively comfortable with the culture of risk as an objective fact (Jasanoff, 1983). However, they were also able to relate to counterarguments about sociopolitical complexity and the often-systemic nature of risk. The findings especially highlight the way practitioners draw discursive resources from both the hard and soft narratives of project risk management as and when required. In this respect, they might be understood as different modes of thinking that operate in parallel (cf. Kahneman, 2011;Slovic et al., 2004). However, abstract cognitive processes are not so easily accessed. They are hence perhaps better understood as modes of narrating in recognition that thinking and doing are inseparable components of practice (cf. Nicolini et al., 2003).
The research participants can also be seen to draw from the emerging narrative of risk governance, at least in terms of emphasizing the importance of providing the necessary information on which company boards can make the necessary decisions. What is clear is that there was no mention of any notion of soft risk management prior to this term being introduced by the research team.

Soft Project Risk Management as an Alternative Narrative
The advocacy of SCA as a framework for uncertainty management was broadly well-received by the research participants. This was especially true of those who did not have a vested interest in existing approaches to project risk management. None of the participants seemed to share Ward and Chapman's (2003, p. 97) concern that SCA "lacks focus on project management issues." In principle, SCA might hence feasibly offer a more nuanced approach to risk governance than that which currently prevails. Indeed, if the aim is simply to ensure that risk is managed responsibly, then the relatively weak reliance on fixed terminologies may well prove to be an advantage (cf. Dowie, 2000). The inherent flexibility of SCA would certainly seem to satisfy the espoused requirement for experiential learning (cf. Boholm et al., 2012).
Perhaps the most notable outcome of the exploration of the SCA was the fleeting acceptance of the merits of allowing participants to express concerns in their own natural language. This was especially striking in the case of Yusuf, who seemingly abandoned his previous dogmatic insistence on use of the 'correct terminology.' However, the extent to which this represented a sustained shift in thinking is questionable. It was more probably a transient concession to fit the mood of the occasion. Ultimately, it would seem that the currently accepted narratives of project risk management are not so easily abandoned.
Even in the emerging space of risk governance any advocacy in support of uncertainty management seemingly challenges the sense of self-identity of those involved. SCA is perhaps more likely to be adopted by generic project managers for the purposes of problem structuring and prioritization. Yet it may also usefully offer a script for practical action without any need to challenge the status of the existing narratives of project risk management (cf. Green, 2001). However, the harsh reality would seem to be that project risk managers talk the language of 'risk as objective fact' because of the way it legitimizes their role. Hence the symbolic enactment of the techniques of quantitative risk analysis has more to do with identity work than with rational choice.

The Geographies of Risk
Of additional significance was Mehmet's emphasis that consideration must also be given to the markets within which Turkish contractors operate. The narratives mobilized by project risk practitioners are not only influenced by the literature, but they are also likely to be influenced by the characteristics of the markets within which they operate. For those operating in regions characterized by instability and conflict, risk is not an abstract probabilistic concept but an essential component of their lived experience. Hence their sense of self-identity is primarily forged from their lived experiences of specific geopolitical contexts rather than the esoteric narratives found in the literature. It was notably those permanently based in central Ankara who aspired most to meeting international standards. This latter group was undoubtedly overrepresented in the workshop and would probably be overrepresented in any such workshop. This is due in part to travel logistics but also to a greater level of tolerance for exploratory events led by academic researchers.
Yet, account must also be taken of the small minority of participants who subsequently suggested a degree of cynicism toward what they perceived as Western methods of project risk management. It was proposed by some that if Turkish firms adhered to the rationalistic methods favored by their Western competitors they would never have ventured into some of their most successful markets. While some may well derive their selfidentities from published best practice narratives of project risk management, others seemingly position themselves in perennial opposition. Such is the nature of identity work.

Conclusion
The essential question that we set out to answer was: Why do the traditional quantitative techniques of project risk management continue to be reproduced in the prescriptive literature if they are so rarely used in practice? Our guiding proposition was that practitioners utilize the accepted narratives of project risk management for the purposes of identity work. The empirical findings provide strong evidence in support of the adopted proposition. There is however no pretense that the narratives generated in the described workshop are representative of any sort of supposed underlying reality. Indeed, their very fluidity is held to be a significant finding in itself. The described empirical narratives are seen to provide indicative examples of the ephemeral processes of identity work. In contrast, the accepted published methodologies of project risk management provide relatively stable scripts from which practitioners derive discursive resources. The published methodologies are hence more usefully understood as being appropriated rather than being implemented. In this respect, the adopted perspective re-conceptualizes the relationship between the theory and practice of project risk management. In essence, it is proposed that narrative is the essential medium through which theory is enacted.
Practitioners are further seen to mobilize discursive resources from different methodologies for the purposes of establishing legitimacy with the various audiences with whom they interact. The resultant unfolding processes of hybridization can be meaningfully understood as identity work. Practitioners however are not only influenced by published methodologies, they also derive stories from their own lived experiences. Such stories are by no means of lesser importance and are often derived from specific geographies.
The proffered analysis stands in stark contrast with the more esoteric debates within the literature whereby different approaches to project risk management are characterized in terms of their underpinning epistemologies. The first such mode of thinking is rooted in the tradition of the scientific method. Hence risk is perceived to be an issue of objective fact and thereby becomes amenable to rational analysis on the basis of probability-impact ratings. The second acclaimed mode of thinking is rooted in the social sciences whereby risk is held to be a social construct. From this perspective, risks are negotiated rather than identified. Hence they become subject to political manipulation. Such debates are not without interest. But the harsh reality is that practitioners tend to be more interested in what works than in issues of epistemology. The essential imperative for practitioners lies in persuading others of their capability in the successful management of risk. Yet the extent of their persuasiveness depends in no small part on the narratives they mobilize. It follows that the advocated modes of thinking are more usefully understood as modes of narrating. Such a view challenges the very notion of methodology as presented in the mainstream project risk management literature.
Despite the highlighted nuances, the dominant narrative of risk as objective fact continues to be persuasive among those predisposed toward rationalistic approaches. This was certainly true of the practitioners involved in the described research. In contrast, the narrative of risk as a social construct found little explicit support. Although the participants listened politely to the idea that risk is socially constructed, they were ultimately more interested in approaches that were supportive of their selfperceived status as custodians of rationality. The narrative of risk as objective fact hence continues to be popular even though its constituent probabilistic techniques are so rarely utilized in practice. There is a long-standing contention that the role of such methods is primarily symbolic. However, to date there has been little focus on how such symbolism contributes to the self-identities of those involved in the enactment of project risk management.
Consideration has also been given to an alternative soft narrative of project risk management as a viable means of legitimizing uncertainty management. The advantage of adopting the advocated narrative of uncertainty is that it is potentially much less restrictive in terms of its embedded terminology. The industry participants responded positively to this suggestion, but were seemingly reluctant to embrace the advocated approach wholeheartedly. The biggest hurdle would again seem to relate to the self-identities of those involved and the narratives they judge will be persuasive in influencing others. The irony is that the continued allegiance to mainstream narratives of project risk management detracts from understanding risk in its more systemic forms. It may well therefore serve as a barrier to effective risk governance.
The described research has further served to highlight the way in which project risk management practitioners frequently find themselves caught between two constituencies. On the one hand, they are frequently responsible for servicing the needs of effective risk governance as administrated by different interests. On the other, they are also often involved in project-level activity working in close interaction with technical specialists. Practitioners are therefore faced with the challenge of continuously maintaining legitimacy in the eyes of different audiences. Different audiences are likely to favor different narratives of project risk management. Hence the tendency for project risk management practitioners to mobilize discursive resources from different formalized narratives for different purposes. Such tendencies are likely to be exacerbated in the case of megaprojects characterized by multiple governance structures, thereby intensifying the requirement for identity work even further. Such an interpretation points toward a radically different alternative research agenda for project risk management. Finally, it is recognized that the arguments presented are likely to be contentious among those who self-identify with existing mainstream narratives of project risk management. It is for this reason that significant care has been taken to engage critically with current theoretical debates prior to offering an alternative perspective based on identity work.