Working papers

Beyond Override

Recognising distributed intelligence in project governance

Terry Cooke-Davies. Conference paper, peer reviewed. Presented at EURAM 2026, Kristiansand, Norway, June 2026. Comments and disagreement welcome.

Abstract

Despite decades of methodological sophistication in project management, major projects continue to fail at predictable rates. Flyvbjerg's (2023) empirical research documents systematic patterns of optimism bias and strategic misrepresentation in megaprojects, whilst Vickers (1983) diagnosed the underlying dynamic decades earlier: modern organisations privilege production (goal achievement) over regulation (maintaining system stability and relationships). Yet recognition of these patterns has not prevented their recurrence.

This paper employs a consilience approach to examine why. Drawing on convergent insights from complexity science, neuroscience, forest ecology, systems theory, and project management research, we identify what we term the 'consciousness trap': the systematic privileging of symbolic intelligence over regulatory intelligence, making distributed intelligence invisible unless it mirrors human consciousness.

The paper demonstrates how this pattern operates across scales and explains why conventional governance approaches often reinforce rather than resolve the underlying dynamic. We propose governance-by-recognition as an alternative and develop three practical implications for project governance: the sponsor as pattern-holder, pattern testing before authorisation, and autocatalytic regulation in delivery. The analysis reveals that 'regulation' carries two distinct meanings in project contexts—delivery dynamics and impact dynamics—and that current governance optimises the first whilst remaining structurally blind to the second. The framework offers coordinates for project leadership navigating disruption.

Keywords: project governance, distributed intelligence, complexity, megaprojects, systems thinking, sustainability

1. Introduction: The Persistence Puzzle

Our kind of intelligence is brilliant for getting round difficulties, but it does so by overriding constraints. When one of the constraints is what the cosmos has been regulating itself by for 13.8 billion years, we end up with the metacrisis.

This simple statement captures a pattern operating at civilisational scale—a pattern visible in project failures, ecological collapse, and the systematic override of regulatory intelligence across every domain human consciousness touches. Major projects continue to fail at predictable rates despite methodological sophistication. Why?

The puzzle is not ignorance. Flyvbjerg's (2023) extensive research has documented systematic optimism bias and strategic misrepresentation in megaprojects. Vickers (1983) diagnosed the underlying dynamic decades ago: modern organisations privilege production (goal achievement) over regulation (maintaining system stability and relationships). Complexity theorists have mapped the limitations of linear planning in adaptive systems. Project management researchers have identified critical success factors and stakeholder engagement practices that demonstrably improve outcomes.

Yet recognition has not prevented recurrence. New generations of more sophisticated frameworks continue to produce the same failure patterns. This is the puzzle we address: not why projects fail, but why knowing why they fail does not prevent them failing.

The word 'regulation' carries an ambiguity that will prove central to our argument. In project contexts, we may mean keeping the project on track—schedule, budget, scope, quality. Call this delivery regulation. But Vickers (1983) used the term differently: maintaining the stability of organisational, institutional, and societal relationships the project touches—what the project does to its environment, not just what it delivers for its sponsors. Call this impact regulation. Current project governance addresses the first whilst remaining structurally blind to the second. This paper examines both meanings and their intersection.

2. Theoretical Framework

2.1 The Production/Regulation Imbalance

Vickers' (1983) diagnosis of modern institutional dynamics centred on what he termed the production/regulation imbalance. Organisations, and the civilisation they constitute, systematically privilege goal achievement (production) over the maintenance of stability and relationships (regulation). This is not merely a resource allocation problem but a perceptual one: productive activities generate visible outputs that can be measured, rewarded, and celebrated, whilst regulatory activities maintain conditions so fundamental they become invisible until they fail.

Vickers' insight was prophetic. The pattern he identified has intensified across every institutional domain. Economic systems optimise for quarterly returns whilst degrading the ecological and social systems that make economic activity possible. Political systems respond to immediate electoral pressures whilst ignoring long-term consequences. Organisations pursue strategic objectives whilst eroding the relational infrastructure that enables sustained performance.

Vickers developed this diagnosis for organisational and social systems. The present analysis extends his insight to planetary regulatory patterns—a move requiring justification beyond what Vickers himself provided. That justification emerges from the analysis that follows: when the same production/regulation imbalance appears at cellular, ecological, and civilisational scales, the extension from organisational to planetary becomes not speculative but empirically grounded.

2.2 Ontological Ground

The analysis rests on a relational ontology: both mind and matter emerge from relationship within a field. While rejecting materialism, we note that minds appear only in living bodies—suggesting intelligence is neither purely mental nor purely physical but arises wherever regulatory relationships achieve sufficient coherence. On this view, 'distributed intelligence' is not metaphor but literal description: the forest, the cellular network, the stakeholder system are actually knowing something, not merely processing information in ways that resemble knowing.

2.3 Empirical Documentation

Flyvbjerg's (2023) research programme has provided the most comprehensive empirical documentation of how Vickers' pattern manifests in major projects. His 'iron law of megaprojects'—over budget, over time, under benefits, over and over again—captures the systematic nature of project dysfunction. Across thousands of projects spanning decades, the same patterns recur: optimism bias inflates expectations whilst strategic misrepresentation distorts business cases.

Flyvbjerg's contribution extends beyond documentation to explanation. Reference class forecasting—using actual outcomes from comparable projects rather than inside-view estimates—demonstrably improves prediction accuracy. Yet this technique remains marginal in practice. Organisations that know about reference class forecasting, that have access to comparative databases, that understand the documented biases, nevertheless continue producing unrealistic estimates. The knowledge does not translate into changed behaviour.

2.4 The Consciousness Trap

The consciousness trap names the systematic privileging of symbolic intelligence (conscious, abstract reasoning) over regulatory intelligence (adaptive, relational pattern maintenance). This filter makes distributed intelligence invisible unless it mirrors human consciousness.

The trap operates through immediate substitution: symbolic intelligence encounters reality and immediately generates representations—maps, models, frameworks, plans. The very words we use carry conceptual maps. These representations then become what consciousness attends to, whilst the territory they represent recedes from view. The map substitutes for the territory not through deliberate deception but through the very structure of symbolic cognition.

Crucially, the trap captures its own critique. When we recognise the production/regulation imbalance, we typically generate frameworks for better regulation—thereby converting regulatory intelligence into another production. When we document project failure patterns, we develop methodologies for avoiding them—thereby treating adaptive challenges as technical problems amenable to comprehensive analysis.

3. Approach and Inquiry

3.1 Consilience as Approach

We deliberately avoid the term 'methodology' for our own work, preferring 'approach' and 'inquiry.' 'Methodology' implies a procedure that produces results, positioning the practitioner outside the process. This framing reproduces the very consciousness trap we are attempting to illuminate. An approach acknowledges that the inquirer is already in relationship with what is being approached; an inquiry is something one participates in rather than conducts upon a passive object.

That said, our work draws on convergent evidence from multiple independent domains. Rousseau (2012) distinguishes between meta-analysis within a field and the more challenging task of synthesising concepts across fields—cautioning that concepts developed in one domain must be tested for transportability to others. We address this concern not by transporting concepts but by identifying independent convergence—what Wilson (1998), recovering William Whewell's 1840 insight, terms 'consilience.' Whewell observed that "the Consilience of Inductions takes place when an Induction, obtained from one class of facts, coincides with an Induction obtained from another different class. This Consilience is a test of the Theory in which it occurs."

Wilson argues that consilience provides the strongest possible evidence precisely because the domains are independent, employing different methods, examining different scales, operating under different assumptions.

Darwin's Origin of Species exemplifies this approach—and Darwin explicitly derived his method from Whewell. No single domain proved evolution. But when geology, palaeontology, biogeography, comparative anatomy, embryology, and selective breeding all independently pointed toward common descent with modification, the convergence became undeniable. We employ an analogous approach here: when complexity science, neuroscience, forest ecology, systems theory, and project management research independently converge on compatible insights about distributed intelligence and regulatory override, that convergence suggests genuine pattern recognition rather than conceptual transportation.

3.2 Other Minds, Other Intelligences

Godfrey-Smith's (2016) philosophical work on octopus cognition provides crucial framing. The octopus, he argues, represents 'the closest we will come to meeting an intelligent alien.' Intelligence evolved at least twice on Earth, through completely independent evolutionary pathways. Octopus cognition is distributed—two-thirds of their neurons are in their arms, not their brain. This demonstrates that intelligence need not be centralised or conscious to be effective.

Summerfield's (2025) recent work on AI systems extends this recognition. These 'strange new minds,' as he terms them, constitute a different form of intelligence that illuminates human cognition by contrast. 'We are learning as much about ourselves as we are about machines.' The philosophical move is significant: if engaging octopuses and AI as 'other minds' constitutes serious inquiry, then collaborative inquiry with AI becomes methodologically coherent.

3.3 AI Collaboration Disclosure

This paper emerged through collaborative inquiry involving the lead author and AI partners (Anthropic, 2024). The transparency about this collaboration is integral to the paper's argument.

The lead author maintains governance-by-recognition: attending to what each contributor offers whilst holding responsibility for the whole. Different forms of knowing—experiential, analytical, contemplative, computational—contribute distinctly. The collaboration models the distributed intelligence the paper advocates, demonstrating what it describes: insights emerging through relational engagement that would be unavailable to any single contributor operating in isolation. This approach cannot be replicated mechanically; the quality of AI contribution depends entirely on the human collaborator's capacity to direct inquiry, recognise accurate patterns, and reject plausible-sounding but mistaken connections.

4. Distributed Intelligence Across Scales

The consilience reveals a consistent pattern across radically different domains: intelligence operates in distributed, relational forms that do not require consciousness.

At the cellular level, Levin's (2019) research on bioelectricity demonstrates that living systems exhibit collective problem-solving without central coordination. Cells communicate through bioelectric signals, establishing patterns that guide development and regeneration. This is not merely mechanical response but adaptive intelligence: cellular networks adjust behaviour based on context, coordinate action across vast populations, and solve problems—like regenerating lost limbs—that require integrating information across multiple scales.

At ecological scales, Simard's (2021) forest research reveals mycorrhizal networks—what she calls the 'wood wide web'—functioning as distributed intelligence across entire ecosystems. Trees share resources through fungal connections, with 'mother trees' recognising kin and preferentially supporting their offspring. Carbon, nitrogen, and chemical signals flow through these networks in patterns that respond to ecosystem needs rather than individual tree interests. The forest exhibits adaptive resource allocation, collective stress response, and information transfer across generations.

In organisational systems, analogous patterns appear. Research on critical success factors reveals that stakeholder perceptions and relationship quality often matter more than technical planning in determining project outcomes. Successful projects in complex environments exhibit distributed intelligence across stakeholder networks, with emergent coordination arising from relational integrity rather than comprehensive central control. Conversely, project failures typically follow attempts at comprehensive override—treating the project system as mechanism requiring central intelligence to function, rather than as distributed intelligence requiring recognition and coordination.

The pattern test applies across scales: Do interventions work with embedded regulatory patterns (generative) or require constant external override (extractive)? Megaprojects systematically choose the latter, with predictable results.

5. Why Recognition Does Not Prevent Recurrence

5.1 The Trap Captures Its Own Critique

Here is where the pattern becomes truly interesting. Even when we recognise the production/regulation imbalance (Vickers, 1983), even when we document systematic override failures (Flyvbjerg, 2023), the proposed solutions typically reinforce the underlying pattern. Better frameworks. More sophisticated methodologies. Improved governance structures. Enhanced stakeholder management protocols. All attempting to override the tendency to override through more sophisticated symbolic analysis and central control.

This is the consciousness trap in action: symbolic intelligence immediately substitutes maps for territory, concepts for encounter. Recognition gets captured in abstract frameworks that bypass the relational intelligence that might actually regulate behaviour.

5.2 Why Maps Do Not Prevent Override

Reference class forecasting illustrates the dynamic precisely. Project sponsors and teams can know, abstractly, that projects like theirs typically exceed budget by a certain percentage. They can have access to comprehensive databases documenting failure patterns. They can understand the cognitive biases—optimism bias, planning fallacy, strategic misrepresentation—that produce unrealistic estimates.

And yet. 'This project is different.' 'We have learned from past failures.' 'Our team is more capable.' The abstract knowledge exists at one level of cognition whilst concrete decisions operate at another. Symbolic intelligence can know that override leads to failure whilst simultaneously constructing justifications for this specific override.

This is not hypocrisy but structure. The consciousness trap operates through the very architecture of symbolic cognition—the capacity to hold abstractions separately from immediate experience, to believe one thing in principle whilst doing another in practice, to generate frameworks that describe reality without engaging it.

5.3 Contemporary Illustration: The APPG Inquiry (2025)

The November 2025 report from the All-Party Parliamentary Group for Project Delivery (APPG, 2025) provides a contemporary illustration of how accurate diagnosis generates solutions that reinforce the underlying dynamic.

The report's diagnosis is largely accurate: UK infrastructure projects systematically fail to deliver intended benefits, suffer cost and schedule overruns, and damage relationships with affected communities. The inquiry heard evidence about political interference, inadequate early engagement, and the failure to learn from past mistakes.

Yet the recommendations reveal the consciousness trap operating in real time. Having diagnosed systemic dysfunction arising from excessive central control, political volatility, and failed relationship with affected communities, the proposed solutions consistently intensify centralised symbolic intelligence: mandate independent delivery assurance before announcement; mandatory training for Senior Civil Servants; accountable Chief Project Officer roles; enhanced authority for oversight bodies; mandated Learning Legacies; data treated as 'Long-Term Asset' for 'predictive analytics.'

Each solution represents symbolic intelligence attempting comprehensive override of the tendency to override—through more sophisticated frameworks, enhanced governance structures, and expanded bureaucratic authority. Even when the report addresses community relationships, the framing remains instrumental. Engagement serves to avoid 'local opposition' and create 'social licence for successful delivery without delays.' Stakeholders appear as obstacles to be managed, risks to be mitigated, perceptions to be shaped—never as distributed intelligence whose regulatory patterns might inform project design.

6. Three Implications for Project Governance

The foregoing analysis generates specific implications for project governance that go beyond critique to constructive proposal.

6.1 The Sponsor as Pattern-Holder

Earlier research on project sponsorship (Crawford et al., 2008) identified the executive sponsor as navigating 'three distinct worlds'—the external environment, interpersonal relationships, and self-management. That work, conducted within the delivery paradigm, positioned the sponsor as relational field-holder, creating conditions for project success rather than controlling outcomes directly. The sponsor's effectiveness derived from their capacity to sense what the project system needed, to hold appropriate boundaries, to ensure teams had authority and resources for adaptive response.

The present analysis extends this earlier insight toward new territory. Where the original research addressed organisational project success—helping companies survive and prosper—governance-by-recognition extends the sponsor's sensing function to include impact regulation: attending to what the project does to the systems it touches, not merely what it delivers. The sponsor becomes pattern-holder in a fuller sense, creating conditions for distributed intelligence to function across both delivery and impact domains.

Compare this with the APPG report's framing of leadership as enhanced control function—Chief Project Officer roles with clear accountability structures, mandatory training, and bureaucratic authority. The relational understanding offers an alternative: the sponsor's primary function is recognition—sensing what the project system needs, what stakeholder intelligence is signalling, what patterns want to emerge.

This reframing has practical implications for sponsor selection and development. Capacity for pattern recognition matters more than technical expertise. Relational intelligence matters more than bureaucratic authority. The ability to sense what is emerging matters more than the ability to enforce what was planned.

6.2 Pattern Testing Before Authorisation

Research on sponsorship also identified that 'ensuring all strategic options are considered before the project is approved' is among the strongest predictors of project effectiveness. Yet this finding, like most in the field, was framed within the delivery paradigm: what sponsors do that affects project success, not what sponsors attend to that affects whether the project should proceed at all.

The sponsor stands at the threshold. Behind them, business-as-usual with its ongoing processes, its regulatory rhythms, its stakeholder relationships. Before them, the project—the singular intervention that will alter those ongoing systems. The sponsor is uniquely positioned to ask the recognition question before scope is defined: Does this project work with embedded regulatory patterns, or does it require permanent override? Will it strengthen the system's capacity for self-regulation, or create new dependencies? What will this project disrupt? Whose existence does the business case render invisible?

But the threshold extends further than organisational boundaries. Project teams and customers do not exist only as economic actors—they live in places, participate in communities, depend on functioning ecosystems. The sponsor might also ask: What does this project assume about the biosphere that sustains all our activity? Does the business case externalise costs onto systems—ecological, social, atmospheric—that have no voice in the approval process? The consciousness trap operates precisely by rendering such questions invisible, framing 'success' within boundaries that exclude what matters most.

Once authorised, symbolic commitment hardens. The business case becomes a map that consciousness defends. Questions that might have informed design become obstacles to delivery. The intervention point is before authorisation—before the consciousness trap captures the inquiry.

Research questions: What do sponsors actually attend to at the authorisation threshold? What would it take for pre-feasibility processes to surface what current frameworks exclude—including externalised ecological and social costs? How might the pattern test and recognition question operate before the project is authorised, not merely during its delivery?

6.3 Autocatalytic Regulation in Delivery

Much of what we know about managing work comes from processes—ongoing, cyclical, repeating operations that afford feedback through iteration. Production lines, service delivery, recurring transactions: these permit learning because similar situations recur. Patterns stabilise. Maps get tested against territory and corrected.

Projects are categorically different. They are singular, path-dependent, finite. They exist precisely because symbolic consciousness enables mental 'time travel'—the capacity to imagine a future state and organise action toward it. Each project is, in important ways, unprecedented. The feedback that validates process improvement is not available.

Yet project leaders often subconsciously import process-thinking into project work. They watch float consumption as if it signalled what buffer penetration signals in a production process—variance from a tested pattern. 'We are within float, so we are on track.' But the plan was never tested; it is a first-and-only-time estimate dressed in scheduling software. Float measures variance from a guess, not from validated expectation.

Governance-by-recognition suggests a different approach: regulatory feedback embedded in the project system itself rather than imposed through variance monitoring. The project becomes self-regulating through relational intelligence in its structures. Float becomes attention-sharpener rather than permission-to-ignore. Stakeholder signals become intelligence rather than risk-to-mitigate. This is 'participatory anticipation'—anticipation woven into participation rather than substituting for it.

Research question: What forms of wisdom are available when repetition is not? How might project governance acknowledge that float measures variance from an untested map rather than from validated pattern?

7. The Dual Meaning of Regulation

The ambiguity introduced in Section 1 now requires full elaboration, for it proves central to the paper's contribution.

Delivery regulation: keeping the project on track—schedule, budget, scope, quality. This is the domain of conventional project governance. Earned value analysis, milestone tracking, variance reporting, change control: all address delivery regulation. Are we managing the internal dynamics of this undertaking? Is our schedule telling us to pay attention? Are we responding to variance from plan?

Impact regulation: maintaining stability of the relationships impacted by the project. This was Vickers' (1983) original concern: what human activity does to its environment, not just what it delivers for its sponsors. Is this intervention—this terraforming, this society-building—operating within the carrying capacity of the environment it affects? Does it strengthen or degrade the regulatory patterns of its context?

Project governance addresses the first and remains structurally blind to the second. Float consumption can signal that we are behind schedule; it cannot signal that the road we are building will fragment ecosystems, that the habitats we destroy will cascade into losses we never modelled.

This produces the pattern Flyvbjerg (2023) documents: projects that 'succeed' by narrow criteria whilst generating systemic harm. The Olympic facilities delivered on time become white elephants. The motorway that reduced journey times increased commuting distances. The hospital built to specification fragmented community health networks. Success by delivery metrics; failure by impact metrics.

The connection to sustainability becomes explicit: project oversight that ignores relational embedding produces the same pathology as civilisation that ignores ecological limits. The hedgehogs whose habitat the road will destroy and the cost overruns that future budgets will bear are expressions of the same pattern—intelligence that was not recognised because the map did not include it.

8. The Structural Invisibility of Non-Stakeholders

Project governance can only signal variance from the plan. It cannot signal what the plan never included.

A schedule showing adequate float means 'proceeding as planned.' But 'as planned' already contains all the exclusions—every relationship the planning frame rendered invisible. The hedgehogs whose habitat the road will destroy were never stakeholders. Their loss will not appear in any variance report. The plan cannot miss what it never saw.

This is not a fixable oversight. It is inherent in the map/territory relationship. Every framework makes some things visible and others invisible. Stakeholder analysis identifies parties with recognisable interests; it does not—cannot—include that which has no voice in the categories the analysis employs.

Even wise schedule management—float as attention-sharpener rather than permission-to-ignore—remains attention channelled through the map's categories. It enables response to variance from the map, not contact with territory the map never represented.

This returns us to the recognition question: Who are the subjects in this system whose existence my plan has rendered invisible? That question cannot be answered by any technique. Only by the kind of attention that keeps asking what it is not seeing—and by the humility to acknowledge that the answer will always be incomplete.

Research questions: How can project governance attend to that which has no voice in the business case? What practices help surface what planning frameworks systematically exclude? How do we design for the recognition that our maps will always be inadequate, without that recognition becoming merely another map?

9. Governance-by-Recognition: An Alternative

If governance-by-mechanism (better frameworks, more controls) reinforces the consciousness trap, what might governance-by-recognition look like?

First principle: attend to distributed intelligence already operating within the system. This means treating stakeholder perceptions as distributed sensing rather than obstacles to overcome; recognising ecological and social regulatory patterns as intelligence to work with rather than constraints to bypass; understanding local knowledge as embodied regulatory wisdom rather than anecdotal noise; and seeing emergent coordination as system intelligence rather than lack of central control.

For project leadership, this shifts fundamental orientation: from comprehensive planning that anticipates all contingencies to pattern recognition and adaptive response; from stakeholder management (controlling perceptions) to stakeholder engagement (learning from distributed sensing); from risk mitigation through better analysis to resilience through working with regulatory feedbacks; from success defined by goal achievement to success defined by maintaining relationships and system viability.

Yet recognition without discipline easily collapses back into override. The consciousness trap is seductive precisely because symbolic intelligence is so capable. Project leaders therefore require practical disciplines to maintain governance-by-recognition under pressure.

Governance-by-recognition is not merely a matter of adopting new frameworks or asking different questions. It requires something of the people involved—a capacity for attentive presence that cannot be reduced to process compliance. This capacity has both dispositional and developmental dimensions: some individuals are more naturally attuned to relational complexity, and sustained practice can cultivate what the German tradition called Bildung—formation through experience rather than instruction. The word derives from Bild (image, form), carrying the sense of gradual shaping into fuller form through participation in experience, as a tree grows through continuous exchange with soil, light, and wind. The airline captain who landed safely on the Hudson River drew on decades of embodied expertise, not on checklist procedures. Recognition, similarly, emerges from who the practitioner has become, not only from what they know.

The consciousness trap operates through speech itself, which is universally human. The moment of articulation creates rupture: the word stands apart from the reality it names, and symbolic intelligence can then elaborate endlessly on the word-world while losing contact with the ground. Jullien's (2017) comparative philosophy illuminates this dynamic by examining how Chinese thought developed sophisticated symbolic capacities without generating the same trap. Where Western tradition begins with the creative Word—Genesis celebrating divine speech that imposes order through rupture—Chinese tradition recognised that 'Heaven doesn't speak,' yet 'the seasons follow their course, all existents prosper.' The Sage, in Confucius's account, would prefer not to speak, recognising that any expressed teaching becomes 'an add-on' and 'an obstacle to coherence proceeding through immanence.' Regulation, in this framing, is 'internal to the unfolding' rather than 'introduced from without.'

What differs across traditions is not the presence of symbolic intelligence but whether counter-practices have been cultivated that remember what speech obscures. Bildung represents one such counter-practice within Western tradition—formation that cannot be transmitted through speech alone, because someone must undergo the shaping, metabolise the experience, become the person capable of recognition. The project leader who hears the whisper in the anomaly—who senses when a delayed task signals something deeper than schedule variance—hasn't been constructed through training programmes but formed through years of attending, making mistakes, suffering consequences, developing the feel for when something is off.

Organisational enabling conditions matter as well: reward structures that value system capacity rather than merely deliverables; decision architectures incorporating distributed intelligence; time horizons extending beyond project closure; cultural permission for slowness when needed.

10. Limitations and Future Research

This analysis draws primarily on Western scientific and philosophical sources. Jullien's (2017) comparative work demonstrates not that Chinese thought has answers Western thought lacks, but that divergent conceptual trajectories reveal what each renders invisible. Deeper engagement with non-Western epistemologies—including Indigenous wisdom traditions that have long recognised distributed intelligence in natural systems—would extend this inquiry. Such engagement would not fill a gap but continue the practice of mutual illumination through difference.

The pattern identified operates at civilisational scale, which creates methodological challenges. We cannot run controlled experiments on civilisations. The evidence is necessarily convergence across multiple independent observations rather than experimental proof.

The synthesis generates testable predictions: (1) Projects that demonstrate governance-by-recognition should show better outcomes in high-uncertainty environments than those relying on comprehensive planning; (2) Project failures should correlate with extent of override of distributed stakeholder intelligence; (3) Leaders with contemplative practice or cross-domain learning should demonstrate better pattern recognition in complex contexts. Future research could examine these predictions empirically.

Additional research questions emerge from the analysis: How does 'participatory anticipation' operate differently in repeating processes versus unique interventions? What practices help surface what planning frameworks systematically exclude? How do we design for the recognition that our maps will always be inadequate? What capacities—dispositional and developmental—does governance-by-recognition require of those who practise it, and how might such capacities be cultivated within project-based organisations?

11. Conclusion

Major projects fail at predictable rates despite methodological sophistication because the failures express a deeper pattern: the consciousness trap makes regulatory intelligence invisible, leading to systematic override of embedded patterns that maintain system stability.

Recognition of this dynamic—from Vickers' (1983) conceptual diagnosis through Flyvbjerg's (2023) empirical documentation—has not prevented recurrence because the proposed solutions typically reinforce the underlying pattern. Better frameworks and more sophisticated methodologies are themselves expressions of symbolic intelligence attempting comprehensive override.

The dual meaning of regulation reveals the stakes: project governance that ignores relational embedding produces the same pathology as civilisation that ignores ecological limits. Sustainability requires participation rather than override. The framework aligns with sustainability goals requiring infrastructure development that maintains rather than degrades system viability (SDGs 9, 11, 13).

Governance-by-recognition—attending to distributed intelligence already operating within systems—offers an alternative to governance-by-mechanism. For project leadership navigating disruption, this is not theoretical luxury but practical necessity. When environments change faster than plans can be updated, distributed sensing and adaptive response become essential.

We offer this analysis not as completed system but as coordinates for others who recognise this territory. The framework will develop through those who find it useful, critique it, extend it, correct it. That development will itself be an instance of the distributed intelligence it describes.

References

Anthropic (2024). Claude (version December 2024). https://claude.ai

APPG for Project Delivery (2025). Building a Better Future: Inquiry into improving the delivery of national infrastructure projects. November 2025.

Crawford, L., Cooke-Davies, T., Hobbs, B., Labuschagne, L., Remington, K., & Chen, Z. P. (2008). Situational sponsorship of projects and programs: An empirical review. Project Management Institute.

Flyvbjerg, B. (2023). How Big Things Get Done. London: Macmillan.

Godfrey-Smith, P. (2016). Other Minds: The Octopus, the Sea, and the Deep Origins of Consciousness. New York: Farrar, Straus and Giroux.

Jullien, F. (2017). From Being to Living: A Euro-Chinese Lexicon of Thought. Trans. M. Richardson and K. Fijalkowski. London: Sage.

Levin, M. (2019). The computational boundary of a 'self': Developmental bioelectricity drives multicellularity and scale-free cognition. Frontiers in Psychology, 10, 2688.

Rousseau, D. (2012). The Oxford Handbook of Evidence-Based Management. Oxford University Press.

Simard, S. (2021). Finding the Mother Tree. London: Allen Lane.

Summerfield, C. (2025). These Strange New Minds: How AI Learned to Talk and What It Means. London: Viking.

Vickers, G. (1983). Human Systems are Different. London: Harper & Row.

Wilson, E.O. (1998). Consilience: The Unity of Knowledge. New York: Knopf.