Programming and self-programming resilience in governance systems: a systems-theoretical investigation

Contemporary public policy operates amid accelerating complexity, polycrisis, and communicative entropy, which undermine forecasting and destabilize established forms of coordination (Cascio et al., 2025; Gots, 2025; Stiegler, 2019). Resilience therefore concerns not merely preparedness for shocks or restoration of a previous state, but the maintenance of coordinated action when actors face novel situations without adequate knowledge or procedures and must devise responses in the course of action. This raises the sociological question of how organizations and interorganizational arrangements preserve operational continuity while transforming the conditions of decision-making.

Existing resilience research identifies relevant resources and adaptive capabilities but explains less precisely the communicative operations through which organizations reproduce and transform themselves. Luhmann's theory of organizations as autopoietic systems of decision communication addresses this gap, while concepts of multifunctionality, meta-organization, and polycentric governance extend the analysis to interorganizational constellations. Because operationally closed systems process crises and policy interventions only by translating them into their own operations, communication channels, and decision premises, resilience-oriented governance cannot directly implement social resilience; it can only shape conditions for endogenous change and governance self-programming.

Resilience plays a dual role in public policy: it is both an objective of interventions seeking to strengthen social adaptive capacity and a mode of governance operation. The distinction between resilience programming and governance self-programming matters because governance arrangements are themselves part of the social systems they seek to influence. Although resilience research often associates polycentric and multilevel governance with adaptive capacity, it says less about how functionally heterogeneous organizations coordinate while preserving autonomy. Combining Teubner's (1993) many-headed hydra with Grothe-Hammer and Rachlitz's (2025) concept of the mingling meta-organization, the article examines how such diversity can support resilience without undermining collective coordination.

The article is theoretical-conceptual in character. The empirical material from Lublin, Poland, is used not as a case study in the strict methodological sense, but as an empirically grounded illustration of the proposed model. The focus is on one urban governance constellation oriented toward the city's social resilience, examined across two areas of public intervention: the European Capital of Culture Lublin 2029 programme and the city's integration policy developed following the refugee crisis caused by Russia's war against Ukraine. The observations draw on material and knowledge accumulated through the author's long-term research on Lublin's integration policy and involvement in ECoC as a researcher and evaluator, providing a longitudinal perspective on local governance processes. The analysis drew on programme documents, participant observation, and attention to the public discourse surrounding both areas of intervention. These areas are treated not as independent cases but as different configurations of the same city-level governance constellation, in which public administration, cultural institutions, NGOs, academic actors, and civic initiatives participate in changing roles and relationships.

The two illustrations trace contrasting pathways of governance self-programming – ECoC Lublin 2029 through the local translation of an externally structured programme, and integration policy through the selective formalization of crisis-induced bottom-up self-organization—showing how the article's theoretical categories become observable in city-level governance.

A theoretical and practical response to contemporary turbulence, visible in both academic literature and policy documents, is the growing use of resilience to describe and program the adaptive capacities of social systems. The term remains conceptually overloaded because it travels across ecology, organizational studies, community research, infrastructure studies, and public policy, carrying different assumptions about what constitutes a disturbance, what the relevant system is, and what kind of continuity should be preserved (Linkov and Trump, 2019). Despite this diversity, most definitions converge around a common conceptual core: resilience refers to the capacity of a system to maintain or reconstitute its identity and core functions under disturbance through some combination of absorption, coping, adaptation, learning, and, where necessary, transformation (Folke, 2006; Walker et al., 2004; Norris et al., 2008; Duchek, 2020). This formulation is useful as an umbrella definition, but it does not imply that resilience is a universal property that can be assessed independently of context. It is better understood as a relation between a particular system architecture, a given disturbance, and the repertoire of responses available to that system.

Holling distinguished resilience from stability: whereas stability concerns the speed of return to equilibrium, resilience denotes the persistence of system identity and relationships through absorption and reorganization (Holling, 1973). Later accounts similarly understand continuity not as restoring the status quo, but as adaptation and, when necessary, transformation; systems may persist by changing structures that no longer support their functioning (Walker et al., 2004). In the social and community domain, resilience is commonly defined as the capacity to cope with disruption, recover, adapt, and, where required, transform. Here, however, the emphasis falls less on the persistence of isolated entities and more on the relationships through which collective responses become possible. Social capital, trust, participation, access to knowledge and resources, and governance capacity shape whether communities can coordinate action and reconstruct their conditions of collective life (Norris et al., 2008; Khan et al., 2024).

Social resilience should consequently not be treated as a homogeneous attribute possessed by a community as a whole. It is a distributed and relational capacity produced through the interplay of multiple actors, organizations, infrastructures, and institutional arrangements (Jungmann and Grothe-Hammer, 2026). This is particularly important for the analysis of cities, whose resilience cannot be reduced to the capacities of municipal administration or any other single organization.

Organizational resilience research addresses a more specific question: how organizations continue operating under surprise and strain, especially when plans fail and established routines no longer provide sufficient guidance. Vogus and Sutcliffe (2007) define organizational resilience as the maintenance of positive adjustment under challenging conditions in such a way that an organization emerges strengthened and more resourceful. Resilience is supported by cognitive, emotional, relational, and structural resources that can be mobilized and recombined in response to unexpected demands. Related work emphasizes practices such as monitoring weak signals, resisting oversimplified interpretations, remaining attentive to failure, learning from near misses, and directing decisions toward those who possess the knowledge most relevant to the problem rather than relying exclusively on formal rank (Weick and Sutcliffe, 2015; Vogus and Sutcliffe, 2007).

Duchek (2020) synthesizes these insights by conceptualizing organizational resilience as a meta-capability comprising three interrelated stages: anticipation, coping, and adaptation. Anticipation includes observing environmental developments, identifying potential threats, and maintaining possible response capacities before a crisis occurs. Coping involves recognizing and interpreting a disruption and devising and implementing responses while events are unfolding. Adaptation comprises reflection, learning, and the incorporation of experience into organizational knowledge, practices, and structures. These stages are analytically distinguishable but neither strictly separated nor linearly ordered. Learning occurs before, during, and after disruption, while previous experience may both enhance preparedness and restrict attention to familiar categories of problems.

These approaches identify the central components of organizational resilience, including monitoring, improvisation, knowledge exchange, social resources, reflection, learning, structural change, and the recombination of available capacities. Yet their broadness leaves an ontological and analytical question partly underspecified: what exactly is an organization, and therefore what must remain continuous, what is disturbed, and through which operations can organizational adaptation occur? Communication, cognition, resources, and structures are often treated as properties of an already constituted collective actor. To make the umbrella definition of resilience more precise for sociological analysis, this article reconstructs these components through Niklas Luhmann's theory of organization. This shifts attention from resilience as a general capacity attributed to a system toward the communicative operations through which an organization constitutes, maintains, and transforms itself.

In Luhmann's theory, an organization is an operationally closed, autopoietic system that reproduces itself through recursively connected communications of decisions. Decisions select among alternatives, refer to previous decisions, and create premises for subsequent ones; the organization persists as long as this connectivity is maintained and its decisions remain attributable to the same system (Luhmann, 2018, pp. 29-35). Organizational continuity therefore does not require unchanged goals, procedures, personnel, or formal structures. Its identity lies in the boundary and historical self-reference that allow it to continue deciding while changing selected structures. This operational perspective accommodates both established bureaucracies and more fluid, hybrid organizational forms (Besio and Tacke, 2024).

Decision-making presupposes uncertainty: decisions are required because several possibilities remain open and the future is undetermined. Uncertainty is thus not an exceptional environmental disturbance but a permanent condition of organizational self-reproduction. Organizations maintain operability through uncertainty absorption: earlier selections are transmitted as premises for later decisions without reopening all the information, doubts, and rejected alternatives from which they arose (Luhmann, 2018, pp. 147-150). Yet each decision also generates new consequences, dependencies, and demands for decisions, so organizational autopoiesis continuously reduces and reproduces uncertainty. This makes Luhmann's approach particularly relevant to resilience in increasingly complex environments: it explains both how organizations achieve sufficient temporary stabilization to act and why established forms of stabilization may need to be reopened when they cease to orient further decisions.

Organizational memory supports this process by selectively distinguishing what should remain available from what can be forgotten. Rather than preserving the organization's complete past, it retains selected decisions, warnings, and response patterns as premises for present action, thereby reducing processing burdens (Luhmann, 2018, pp. 123-127, 152-155). Remembering and forgetting are both resilience-relevant: memory preserves identity and reusable experience, while forgetting releases processing capacity. Reusing the past is itself a new decision about its relevance, and excluded alternatives or unsuccessful experiments may be reactivated when circumstances change. Archives and databases contribute to resilience only when their contents can re-enter decision communication.

Decision premises provide the structural form of uncertainty absorption. Luhmann distinguishes programmes, personnel, and communication channels: they respectively specify criteria for correct decisions, connect expectations to particular persons, and regulate participation and the circulation of binding communication (Luhmann, 2018, pp. 181-193). These decidable premises coexist with historically formed conditions—including culture, trust, tacit expectations, professional identities, and established definitions of problems and responsibility—that cannot be redesigned directly. They can be influenced only indirectly through changes in formal premises, recurrent interactions, participation, and the processing of conflict and experience. Resilience cannot therefore be programmed solely through formal plans and procedures.

Operational closure does not imply isolation. Legal changes, crises, research findings, stakeholder expectations, and the actions of other organizations become consequential only when reconstructed through the organization's own distinctions as information connectable to decisions (Luhmann, 2018, pp. 38-51). The same event may consequently be treated as a crisis, a routine problem, or remain unnoticed by different organizations. Resilience requires sustained irritability rather than maximum sensitivity: the selective capacity to register discrepancies between expectations and events and to reopen uncertainty when established assumptions fail (Luhmann, 2018, pp. 167-171). Adaptation is therefore an internally generated process of learning and structural revision, not the direct importation of environmental solutions.

Organizational resilience should not, however, be equated with survival. It concerns the more specific capacity to preserve the connectability of decisions under intensified uncertainty, surprise, and the failure of established expectations, particularly when this requires modifying the structures that previously supported decision-making. Insufficient change may prevent the organization from processing novel conditions, whereas indiscriminate change may dissolve the memory, distinctions, and premises on which further decisions depend. Continuity is achieved through selective structural revision rather than either persistence or transformation alone.

This reconstruction incorporates within one communicative framework the monitoring, coping, learning, and reflexivity emphasized in organizational resilience research (Vogus and Sutcliffe, 2007; Weick and Sutcliffe, 2015; Duchek, 2020), while adding a more precise account of uncertainty, organizational memory, decision premises, operational closure, and self-reproduction. Anticipation concerns decisions about what to observe, which weak signals to treat as information, and what capacities or experiences to retain. Coping consists in provisional uncertainty absorption through problem redefinition, resource recombination, altered responsibilities, and improvisation. Adaptation involves observing consequences, learning and forgetting, revising decision premises, and restabilizing useful variations. Duchek's (2020) three dimensions can thus be retained if understood as recursively connected modes of decision-making rather than a linear sequence. Communication is constitutive here: knowledge, expertise, trust, and resources support resilience only when they enter relevant communication channels and alter subsequent decisions; otherwise, even formally recorded “lessons learned” do not constitute organizational learning.

Reflexivity denotes an organization's capacity to make the conditions of decision-making objects of further decisions, although complete self-knowledge remains impossible because the system is opaque to itself. Self-programming operationalizes this capacity by revising formal decision premises such as programmes, roles, communication channels, and participation rules. Trust, professional identities, informal expectations, and organizational culture can be affected only indirectly, so resilience-oriented change must address both formal premises and the historical conditions of their connectability. Structural change therefore proceeds less through the implementation of a complete design than through variation, selection, and restabilization: improvisations are retained, modified, or forgotten. Loose couplings may contain disturbances and enable experimentation, but can also insulate situated practices from centrally initiated reforms (Luhmann, 2018, pp. 273-291).

Organizational resilience can thus be defined as the capacity of an operationally closed yet environmentally irritable system of decision communication to preserve decision connectability under disturbance by selectively absorbing and reopening uncertainty, revising decision premises and coordination, learning and forgetting, and restabilizing structures without losing operational continuity and identity.

The definition is analytical rather than normative. An organization may maintain and adapt its decision-making while reproducing harmful goals, inequalities, or dysfunctional practices. Organizational resilience must therefore be distinguished both from the desirability of organizational outcomes and from wider social resilience, which depends on relations among organizations, residents, infrastructures, and governance arrangements rather than on the adaptability of any single organization.

The concept of multifunctional organization specifies an important source of organizational resilience: the capacity to process expectations associated with different functional domains. Roth and Sales (2025) challenge the assumption that organizations can be assigned exclusively to a single function system, treating multifunctionality as a normal structural condition and strict monofunctionality as a limiting case. Because organizations remain operationally closed, they selectively translate political, legal, economic, scientific, educational, or cultural expectations into their own programmes, roles, indicators, budgets, and decision criteria. Under uncertainty, this capacity can enhance resilience by increasing observational variety, enabling the recombination of competencies and resources for unforeseen tasks, and facilitating translation between organizations with different primary orientations. Cultural organizations, universities, and NGOs can consequently redeploy capacities developed for artistic, scientific, educational, service, or advocacy purposes in response to emerging public problems.

The resilience contribution of multifunctionality nevertheless depends on selectivity. Incompatible expectations may generate overload, mission drift, and conflicts over organizational identity, while the domination of one rationality may suppress the differentiated observations and competencies that provide adaptive variety. Multifunctionality must therefore be stabilized through decision premises specifying which functional references are relevant, how conflicts among them are processed, who performs translation, and how its results enter organizational memory. Through self-programming, organizations can expand, restrict, or reconfigure their functional repertoires while preserving operational coherence and the possibility of co-evolution among distinct systems and organizations (Van Assche et al., 2022). Multifunctionality should thus be treated—alongside uncertainty and the processual autopoiesis of organization—as a structural condition of resilience: it increases adaptive capacity only insofar as heterogeneous expectations remain connectable to decision communication.

Having reconstructed organizational resilience as the preservation of decision connectability through revisable premises, the analysis now moves beyond the individual organization. Contemporary crises—from migration and public-health emergencies to climate disruption and social polarization—cross organizational, policy, and functional boundaries. Responding to them requires political, legal, scientific, administrative, economic, cultural, and civic capacities distributed among autonomous organizations with different programmes, time horizons, criteria of success, and problem definitions.

Networked, multilevel, and polycentric governance therefore appear promising under uncertainty. Multiple, partly overlapping decision centres can support experimentation, contextualized responses, learning, access to diverse knowledge, and risk distribution (Folke et al., 2005; Ostrom, 2010). Redundancy may prevent one unit's failure from disabling the whole, while local autonomy shortens the distance between observation and intervention. Adaptive-governance research similarly emphasizes self-organization, bridging organizations, social memory, and the recombination of actors and resources (Folke et al., 2005).

Yet polycentricity is not inherently resilient. The same features may produce fragmentation, incompatible standards, coordination costs, unequal participation, decision blockages, and diffuse accountability. Neither more decision centres nor denser coordination necessarily improve collective action. The relevant question is whether governance can revise relations among centralized and decentralized, tightly and loosely coupled, or formal and informal modes when disturbances invalidate established divisions of observation, decision-making, and responsibility—and whether autonomy and functional heterogeneity enable or impede cooperation.

Conceptual caution is therefore required. In Luhmannian terms, a social system exists only where communication recursively reproduces a system-environment boundary; not every interorganizational arrangement qualifies. I use governance constellation for organizations selectively coordinating around a problem and reserve governance system for arrangements reproducing an identifiable communicative order through recurrent operations, expectations, roles, memory, and boundaries. The resilience of participating organizations, the interorganizational order, and the wider community remain distinct but interdependent. Governance resilience lies in relations that make distributed capacities available for collective problem-processing, not in organizational survival or network density alone.

Teubner's network theory explains how an interorganizational arrangement can develop capacities not reducible to those of its participants. Although formulated primarily in relation to economic and contractual networks, it can be adapted to public governance through three mechanisms: double attribution, the combination of variety and redundancy, and multistability (Teubner, 1993, pp. 41-60). A network becomes an emergent order when actions are recursively attributed both to the organization performing them and to the wider arrangement. They remain subject to the organization's programmes and responsibilities while also being recognized as contributions to shared commitments. When subsequent operations refer to them in both respects, recurrent cooperation produces network-level expectations, memory, and collective agency without creating a sovereign decision-making centre – the form captured by Teubner's metaphor of the “many-headed hydra” (Teubner, 1993, pp. 49-50, 57).

From a resilience perspective, double attribution enables locally developed responses to become resources for the wider constellation. An adaptation remains the decision of a particular organization, but it may inform action elsewhere when it is communicated, recognized, and retained in subsequent network operations. Autonomous organizations provide variety through their different knowledge, competencies, environmental relations, and response strategies, while recurrent relations create redundancy by stabilizing shared expectations and accumulated experience (Teubner, 1993, pp. 48-50). Multistability further allows responses to shift among individual organizations, subsets of participants, and the wider network as disturbances evolve. Network resilience therefore lies neither in maximum decentralization nor in complete integration, but in making distributed capacities collectively accessible while preserving the autonomy required for experimentation and alternative responses (Teubner, 1993, p. 51).

Urban governance is nevertheless more heterogeneous than Teubner's paradigmatic economic networks. It connects public authorities, NGOs, universities, firms, and cultural institutions that process political, legal, economic, scientific, and cultural expectations through different decision premises. These organizations may be structurally multifunctional, but their processing of functional expectations remains selective (Roth and Sales, 2025). Teubner's theory explains how their actions can acquire a network-level attribution, but not how knowledge, coordination, representation, and responsibility can be formally stabilized across such a cross-domain arrangement. Meta-organization theory addresses this second problem.

A meta-organization is a formal organization whose members are organizations and which possesses its own membership boundary, decision processes, identity, and organizational address (Grothe-Hammer and Rachlitz, 2025, p. 557). It creates two autonomous but interrelated levels of decision communication: a meta-level decision can become a premise within a member organization, while members' decisions may be incorporated into the meta-organization's operations. This arrangement can stabilize communication and institutional memory, establish common roles and rules, process conflicts, represent the constellation externally, and make selected decisions and responsibilities organizationally addressable. It may also generate disputes over competence, autonomy, representation, and accountability (Grothe-Hammer and Rachlitz, 2025, pp. 557-560).

Grothe-Hammer and Rachlitz's concept of the mingling meta-organization is particularly relevant because it describes organizations whose members have different primary functional orientations but cooperate around a common issue. It shows how heterogeneous perspectives can enter a formal decision process, although often at the cost of narrowing the agenda to a provisional common denominator (Grothe-Hammer and Rachlitz, 2025, pp. 567-568). The category can be applied to urban governance only by analogy, however. Problems such as social resilience, integration, and urban regeneration cut across functional domains, while governance constellations also include residents, informal initiatives, temporary partnerships, and spontaneous forms of self-organization that cannot be incorporated fully into a formal organization of organizations.

The resulting model is therefore a hydra-like governance constellation with a meta-organizational or quasi-meta-organizational core. The network preserves distributed agency, organizational variety, peripheral innovation, and multiple points of environmental contact, while the core formally anchors selected capacities for memory, coordination, representation, conflict processing, and responsibility. It cannot decide for the entire constellation or determine the internal operations of autonomous participants; its contribution depends on remaining open to knowledge and initiatives arising beyond its formal boundaries. Teubner attributes collective actorhood to emergent networks through double attribution, whereas meta-organization theory reserves full decision-making autonomy, addressability, and responsibility for formal organizations. Used complementarily rather than ontologically synthesized, these perspectives explain how distributed network agency can coexist with a formal core that stabilizes capacities an emergent network may otherwise lack.

The central design problem of resilient governance lies in balancing autonomy and coordination. Autonomy preserves distributed observation, specialization, local initiative, and experimentation, whereas coordination enables participants to construct a shared situational picture, mobilize complementary resources, manage interdependencies, and allocate responsibility. Too little coordination produces fragmentation, incompatible interventions, and responsibility gaps; too much creates bottlenecks, suppresses local knowledge, increases dependence on central actors, and may eliminate the differences from which adaptive variety arises. This relation should therefore be treated not as a fixed optimum but as an object of observation and revision. Resilience lies in selectively recoupling and decoupling organizations: strengthening coordination where joint stabilization is required, loosening it where experimentation is more valuable, establishing temporary couplings around particular disturbances, and redesigning relations that have become rigid or counterproductive. This co-evolutionary perspective shifts attention from an ideal governance architecture to the revisability of the relations through which knowledge, resources, expectations, and decisions circulate (Van Assche et al., 2022).

At this level, we propose using the term metacoordination to describe the organization of conditions under which multiple processes and modes of coordination can take one another into account—a coordination of coordination. In Luhmannian terms, a coordinating organization may revise its own programmes, roles, procedures, and communication channels and thereby make certain decisions by other organizations more probable, but it cannot determine how autonomous organizations reconstruct these conditions within their own decision contexts. This formulation converges with metagovernance theory's emphasis on reflexively combining coordination modes without restoring a unitary governmental centre (Jessop, 2003; Sørensen and Torfing, 2009). Metacoordination is supported by communication and translation infrastructures, procedures for joint action and conflict processing, and feedback mechanisms that make coordination arrangements observable and revisable. It differs from meta-organization: the latter is a formal organization with organizational members, whereas metacoordination is a function that may be formally anchored or distributed across the governance constellation. A meta-organization may support this function but does not guarantee it, since it may privilege dominant members or prevent peripheral knowledge from entering collective decisions. Metacoordination must therefore remain reflexive both about the adequacy of the current combination of hierarchical, negotiated, networked, and self-organized coordination and about its own selectivity. Participation rules, information standards, interfaces, and problem definitions inevitably privilege some actors and observations; reflexivity cannot eliminate this selectivity but can make its consequences visible, contestable, and revisable.

One practical implication is that reflexivity may be supported by evaluation understood as part of governance's communicative infrastructure rather than merely as ex post measurement or accountability. Evaluation can function as selective meta-observation by reconstructing how participants define problems, interpret actions, attribute responsibility, and understand relations between organizational contributions and collective effects. It thereby offers a necessarily partial self-description of the governance constellation, making causal assumptions, interdependencies, exclusions, and coordination failures available for communication. Evaluation contributes to resilience, however, only when its observations enter decision communication, are retained in institutional memory, and prompt revisions of decision premises and coordination arrangements. It provides no neutral external view but can strengthen the system's capacity for self-observation and transformation.

Resilience increasingly serves as a programmatic orientation in public policy, defining desirable forms of continuity, adaptation, and transformation under uncertainty (Linkov and Trump, 2019). In EU cultural policy, culture is correspondingly framed as a resource for cohesion, democratic robustness, participation, well-being, and adaptation, while the Culture Compass for Europe links these objectives to expanded monitoring and policy-learning capacities (European Commission, 2025). This may be interpreted as a gradual reprogramming of EU cultural governance through soft-law instruments and supporting competences rather than direct harmonization (Dorotić, 2023). A systems-theoretical perspective, however, challenges the assumption that such objectives can be implemented through ready-made models introduced into cities or sectors from above. Because organizations and other social systems are operationally closed, external interventions can only generate irritations that become consequential when their addressees select and translate them into internally connectable communications and decision premises (Luhmann, 2018, pp. 33-35, 49-51).

Operational closure therefore grounds the concept of self-programming: EU policies may provide objectives, categories, resources, and incentives, but local governance must reconstruct them as locally meaningful programmes, roles, relationships, and modes of coordination. Similarly, urban resilience emerges through changes in locally embedded communications, practices, relationships, and collective action rather than through the programming of a unitary community. Resilience programming should therefore be understood as reflexively shaping conditions for endogenous social change and governance self-programming, not as implementing a predetermined external model. In the next section, we draw on examples of policies aimed at strengthening resilience to illustrate the usefulness of the understanding of resilience developed in this article.

The European Capital of Culture (ECoC) framework seeks to use culture to support long-term urban development, increase cultural participation, promote inclusion, and strengthen relations within and beyond the cultural sector. Although resilience is not always explicitly defined in ECoC policy, the programme incorporates many of its dimensions, including social, organizational, economic, psychological, and ecological capacities. Resilience thus operates as an umbrella semantics connecting different policy objectives with the broader capacity of cities and their cultural sectors to respond to disruption and cumulative stress. At the same time, the EU framework leaves candidate cities substantial freedom to contextualize its general objectives. The policy architecture behind ECoC therefore appears to recognize that common programme assumptions require local translation rather than uniform implementation (European Parliament and Council, 2014).

ECoC Lublin 2029 translates this orientation into a project of cultural and social regeneration responding particularly to the consequences of the COVID-19 pandemic, the disruption of social relations, and the migration crisis caused by Russia's full-scale invasion of Ukraine. Its central idea, RE:UNION, refers to renewed integration, the repair of broken ties, and intergenerational, cross-sectoral, and intercultural dialogue. The programme is presented as a long-term process intended to strengthen the city's capacity for cooperation, mutual care, and adaptation (ECoC Lublin 2029, 2024). It is embedded in Lublin's wider cultural and development policy and can therefore be understood as its temporary intensification and partial reprogramming rather than as an isolated intervention ending in 2029.

The successful application was based on extensive bottom-up involvement by Lublin's heterogeneous cultural environment. Local organizations, artists, informal initiatives, experts, and municipal actors participated in translating general EU and ECoC objectives into locally meaningful problems, priorities, and projects. This participatory process did more than supply the application with content. It enabled participants to recognize the programme as partly their own undertaking, linking their organizational or professional activities with the wider ECoC project. In Teubner's terms, it created conditions for double attribution: contributions could be understood simultaneously as the actions of autonomous participants and as operations of a broader interorganizational constellation.

The programme's theory of change connects three distinct but interdependent dimensions of resilience. Governance resilience rests on inclusive and participatory implementation that mobilizes heterogeneous knowledge and resources, preserves autonomous initiative, distributes responsibility, and retains the capacity for programme-wide decisions. Resilience thus denotes both an intended effect and a mode of operating ECoC. Cultural-sector resilience is pursued through local, cross-sectoral, and international networks that expand access to competencies, resources, partners, and alternative responses. It encompasses financial, infrastructural, relational, competence-based, and well-being capacities and, in systems-theoretical terms, the ability of cultural organizations to continue and revise decision-making under turbulence. Its durability depends on whether partnerships and solutions developed through ECoC enter organizational and interorganizational memory. Urban social resilience, in turn, is pursued through cultural participation and co-creation: shared authorship is expected to foster identification, ownership, co-responsibility, trust, bridging ties, and collective action across social boundaries.

These dimensions form a potentially self-reinforcing rather than linear process. Resilient governance can strengthen the cultural sector; a more connected sector can expand participatory cultural activity; and residents' engagement can return knowledge, legitimacy, and resources to both the sector and the governance constellation. These relations express the programme's theory of change, not an empirically established causal sequence. Their circularity nevertheless shows why resilience cannot be installed as a discrete outcome: it depends on mutually conditioning processes through which organizations, networks, and communities reproduce and transform their own capacities.

The early implementation of ECoC illustrates the tensions accompanying the transition from the consortium-like arrangement that prepared the application to a formal institution responsible for implementation. The former involved municipal and cultural institutions, NGOs, artists, informal initiatives, experts, and local communities, whereas institutionalization concentrated the organizational address of decision-making and responsibility. Some initial choices—including the selection of the coordinating institution's leadership and decisions concerning particular artistic content—were interpreted by parts of the cultural environment as external and top-down, and thus as departing from the declared principles of participation and co-creation. These perceptions were accompanied by negative reactions, selective demobilization, contestation, and challenges to the legitimacy of particular decisions, temporarily and unevenly weakening some actors' sense of agency and identification with the programme. Public and media discussions also revealed how stakeholder diversity complicates the communication of programme priorities, decision premises, and responsibilities.

These tensions demonstrate that externally formulated objectives become effective only when local actors reconstruct them through their own identities, distinctions, programmes, and decision premises (Luhmann, 2018, pp. 38-51). The bottom-up preparation of the application facilitated this reconstruction and supported double attribution, allowing actors to understand their autonomous activities as contributions to a shared programme; conversely, decisions perceived as disconnected from that process risked shifting attribution toward the formal centre and weakening mobilization. Formal institutionalization nevertheless remains necessary to preserve programme memory, allocate resources, represent ECoC externally, process conflicts, and make decisions and responsibilities addressable. The challenge is therefore to combine a formal coordinating core with distributed agency through stable participatory and communicative interfaces that keep heterogeneous contributions connected to programme-level decisions. Understood as the core of a broader, hydra-like constellation, the coordinating institution can support collective orientation without absorbing the network's autonomy or peripheral experimentation (Teubner, 1993; Grothe-Hammer and Rachlitz, 2025). This requires decisions to be not only formally attributable but also communicated in ways that make their objectives, rationale, and premises intelligible across the constellation.

The theoretical framework developed above helps explain these tensions. Participants observe ECoC through different organizational roles, time horizons, criteria of success, and understandings of the programme, while fragmented access to one another's interpretations can generate misunderstanding and mistrust. Since this diversity cannot be eliminated without weakening the programme's adaptive potential, the relevant challenge is metacoordination: revising how initiative, coordination, and responsibility are distributed. ECoC may require concentrated decision-making in some situations and decentralized experimentation, consultation, or temporary task arrangements in others. Its multistability depends on the ability to move between these modes when existing arrangements become inadequate.

Communication is therefore central to governance self-programming. Translation interfaces should make decision premises and disagreements accessible, enable feedback, and demonstrate how peripheral contributions affect programme-level decisions. This supports double attribution by allowing participants to retain their autonomy while identifying their activities as contributions to ECoC. Evaluation, already an important component of the programme, may reinforce this capacity if it encompasses implementation, interorganizational relations, and the changing environment rather than merely assessing outcomes ex post. Because participating actors define programme success through different perspectives and criteria, a standardized evaluative framework is insufficient; evaluation must therefore be participatory and make these definitions visible and discussable. Its value lies in introducing such observations into communication and decision processes so that tensions, assumptions, and coordination arrangements can become objects of collective learning and revision.

The ECoC illustration shows that resilience operates simultaneously as an intended societal effect and a mode of governance, neither of which follows automatically from cultural projects, participation, formal coordination, or network density. Programme effects emerge recursively from the self-programming of the governance constellation, the adaptive capacities of cultural organizations, and residents' participation and social relationships—all internal to the urban society whose self-transformation ECoC seeks to stimulate. Operational closure explains why general objectives require local reconstruction, while double attribution, multistability, meta-organization, and metacoordination indicate how this process may become collective and durable. The relevant problem is therefore not choosing between hierarchy and networks but maintaining a revisable relation between them: formalization provides memory, accountability, and decision capacity but may weaken identification by displacing distributed agency, whereas participation mobilizes heterogeneous capacities without removing the need for selection, responsibility, and conflict processing. Resilience-oriented self-programming consists in keeping these tensions observable and revisable through communication, translation, learning, evaluation, and flexible distributions of initiative and responsibility, thereby enabling collective action without suppressing the autonomy and diversity on which adaptive capacity depends.

Whereas ECoC begins with an externally structured programme requiring local translation, Lublin's integration policy followed an almost opposite trajectory. The rapid arrival of people displaced by Russia's invasion of Ukraine in 2022 exceeded existing administrative structures and procedures. The response emerged through parallel decisions by public institutions, NGOs, businesses, universities, religious organizations, informal groups, and residents. Building networks was therefore a coping strategy from the outset: organizational variety expanded the available knowledge, resources, and responses, while overlapping assistance and communication channels provided redundancy when no single actor could fully observe or process the crisis (Stawicki and Dziekanowska, 2025).

The Lublin Social Committee to Aid Ukraine exemplified this emergent capacity. Its relatively flat, task-oriented structure combined diverse competencies through direct communication, trust, and flexible roles, directing decisions toward actors with the resources most relevant to particular problems rather than those holding the highest formal positions (Podgórska et al., 2024). The response also enabled bottom-up learning: a crisis-management model developed by one local organization was subsequently adapted at the regional level (Stawicki and Dziekanowska, 2025). The solution remained attributable to its organizational originator while becoming part of the wider support system's memory. Responses could thus be developed autonomously, coordinated through the Committee, or undertaken by municipal and regional administration, illustrating multistability rather than simple decentralization.

As the emergency stabilized, governance shifted from generating improvised solutions to selecting and retaining them within longer-term integration policy. Formalization was necessary for continuity, funding, legal accountability, and equal access to services, but it also altered the relations that had enabled rapid adaptation. Excessively tight coupling risked subordinating civic initiatives to administrative procedures and suppressing peripheral knowledge, whereas loose coupling could leave cooperation dependent on temporary projects and personal relations. The resilience challenge was therefore to formalize selected solutions without eliminating the self-organized capacities from which they emerged.

Integration and migration policy nevertheless remain highly exposed to turbulence arising from the continuing war in Ukraine, changing migration patterns and legal frameworks, unstable funding, polarized political narratives, and disinformation. Under these conditions, partnerships among municipal institutions, NGOs, researchers, businesses, and migrant organizations perform a broader resilience function. They support the production and cross-validation of sufficiently credible knowledge to guide decisions and counter interpretations shaped by disinformation or populist politics. International and inter-municipal networks additionally provide comparative experience, institutional partners, and alternative sources of expertise and funding, reducing dependence on unstable local or national arrangements.

Networked knowledge production does not create a single authoritative epistemology. Scientific findings can be compared with administrative data, practices adopted in other cities, the experience of organizations working with migrants, and migrants' own perspectives. Conversely, externally identified good practices can be assessed through local research and dialogue. This observational redundancy helps reveal inconsistencies, blind spots, and contextual limitations, making policy less dependent on any single source of knowledge.

Research and recurrent knowledge exchange therefore do more than circulate information. They enable the integration-policy field to observe how its participants define problems, select evidence, assess success, and attribute responsibility. In this sense, they contribute to the self-description of the governance constellation and create an infrastructure for reflexive metacoordination: heterogeneous observations can be compared and translated, while existing coordination arrangements become open to revision. Learning occurs when these observations enter organizational decision processes and modify programmes, responsibilities, participation arrangements, or communication channels; otherwise, they remain external information rather than elements of governance self-programming.

The illustration also shows why organizational and governance resilience cannot be equated. An organization may protect its continuity by withdrawing from cooperation or narrowing its responsibilities, while the interorganizational arrangement depends on making organizational capacities available for collective action without abolishing their autonomy. Its resilience therefore requires sufficient trust, shared memory, translation, and identifiable responsibility to allow participants to communicate uncertainty and failure, learn from one another, and revise their relations under changing conditions.

This article speaks to interdisciplinary resilience research that conceptualizes resilience through continuity, coping, learning, adaptation, and transformation (Holling, 1973; Walker et al., 2004; Norris et al., 2008; Vogus and Sutcliffe, 2007; Duchek, 2020), as well as to adaptive-governance research emphasizing polycentricity, self-organization, and distributed knowledge (Folke et al., 2005; Ostrom, 2010). These literatures identify important resources and capacities but leave less explicit how resilience is communicatively constituted, how organizational capacities become available at the governance level, and how governance arrangements can reflexively transform their own modes of coordination. The article addresses this gap by linking organizational reproduction, interorganizational coordination, and endogenous social change while keeping organizational, governance, and community resilience analytically distinct.

The theoretical synthesis combines Luhmann's account of decision communication and operational closure (2018) with Roth and Sales's concept of multifunctionality (2025), Teubner's theory of emergent network agency (1993), and Grothe-Hammer and Rachlitz's analysis of meta-organizational anchoring (2025). This combination explains how structurally multifunctional organizations can enter co-evolutionary relations stabilized by reflexive metacoordination, making their distributed observations, resources, and adaptations collectively usable while formally anchoring selected functions of memory, representation, and responsibility. Organizational and functional differences are both a source of observational variety and a potential cause of fragmentation; governance resilience therefore depends on keeping these differences and the couplings among organizations observable and revisable rather than maximizing either autonomy or integration.

This argument also contributes to debates on policy implementation, policy learning, and evidence-based policy. Operational closure challenges linear models in which policy objectives or evidence are transferred into passive target systems: interventions become consequential only through endogenous reconstruction. Resilience programming should therefore shape communicative and organizational conditions for adaptation, while governance self-programming consists in revising decision premises, interfaces, participation rules, memory, responsibilities, and coordination modes. This complements complexity-oriented critiques of instrumental policy making (Sanderson, 2009) and extends accounts of evaluation as a mechanism of policy learning (Sanderson, 2002). From the perspective developed here, evaluation should not be confined to ex post assessment of outcomes; organized as a process-wide communicative infrastructure, it can support higher-order observation and the self-description of a governance constellation, making its causal assumptions, exclusions, and coordination relations available for revision.

The Lublin illustrations make these propositions visible without testing them causally. ECoC Lublin 2029 shows how externally formulated objectives require local translation and how societal resilience as a policy goal depends on resilience as a mode of governance. Integration policy shows the reverse movement from distributed crisis improvisation toward the selection and formal anchoring of useful solutions. Together, they indicate that externally initiated programmes and bottom-up self-organization become durable through recursive processes of translation, selection, learning, and restabilization.

This approach does not assume that resilience is inherently desirable. Organizations and governance arrangements may resiliently reproduce exclusion, domination, or harmful objectives, while participation, self-organization, and multifunctionality may reinforce unequal epistemic authority. Normative assessment must therefore ask what is being made resilient, for whom, against which disturbances, and at whose cost. It must also examine who can define collective problems, whose observations enter collective memory, which failures can be communicated without sanction, and who controls the interfaces through which knowledge and resources circulate.

The framework remains a theoretical reconstruction supported by illustrations rather than an empirical assessment of resilience outcomes. Further research should determine when a governance constellation develops sufficiently stable identity, memory, expectations, and boundaries to be treated as an emergent system and should operationalize double attribution, multistability, translation capacity, systemic memory, and reflexive self-programming. Longitudinal and comparative studies could examine when formalization retains rather than suppresses peripheral innovations, how meta-organizational cores affect organizational autonomy, and whether governance arrangements can deliberately revise relations among hierarchical, negotiated, and self-organized coordination. They should also investigate how the benefits and costs of governance resilience are socially distributed. The broader claim is that social resilience can be strengthened neither by imposing predefined solutions nor by relying on spontaneous self-organization alone, but by enabling autonomous systems to transform themselves while reflexive metacoordination keeps their distributed capacities available for collective action under uncertainty.