What Kind of Knowledge Does the CCF Produce? A Critical Realist Account of Craft Enterprise Management Science
Management science has an epistemological problem it rarely discusses in public. It borrows its quantitative methods from positivism — the philosophy of science that treats social phenomena as if they obey the same kind of causal laws as natural phenomena, and that evaluates knowledge claims by their predictive power across large samples. It borrows its qualitative methods from interpretivism — the philosophy that treats social phenomena as constituted by meaning rather than governed by causal laws, and that evaluates knowledge claims by their capacity to illuminate the subjective experience of participants. And it routinely combines both traditions in the same research programme without specifying which standards of evidence apply to which claims.
The Coetzee Convergence Framework has this problem. It relies on meta-analytic effect sizes drawn from positivist research programmes — Seibert, Wang, and Courtright’s (2011) ρ = .44 is the product of a quantitative research tradition that treats psychological empowerment as a variable whose causal effects can be estimated across independent samples. It also acknowledges that the founder’s identity meaning-making determines whether the Safety Pre-Flight protocol produces genuine psychological safety or compliance theatre — a claim that is interpretivist in character, because it insists that the causal mechanism (safety) is constituted by the founder’s subjective experience of the intervention, not merely by the intervention’s objective structure. And it proposes emergent organisational properties — the Verified Empowered Routine — that are neither purely objective (they depend on participants’ meaning-making) nor purely constructed (they have real causal powers that exist independently of any particular participant’s beliefs about them).
A framework that spans these three epistemological traditions without specifying its commitments is a framework that cannot be fully evaluated, because the standards by which its claims should be assessed depend on the philosophy of science that underlies them. This article specifies those commitments. It does not resolve the epistemological tensions in management science — they are not resolvable within a single article or a single framework. What it does is locate the CCF within the tradition that best accounts for its specific combination of claims, and specify what that location implies for how the framework’s evidence programme should be conducted and evaluated.
Critical Realism as the CCF’s Epistemological Home
Roy Bhaskar’s (1975) A Realist Theory of Science introduced the philosophical framework that has come to be known as critical realism through a deceptively simple question: what must the world be like for science to be possible? Bhaskar’s answer was that the world must be stratified — composed of structures and mechanisms at multiple levels of reality, each with genuine causal powers, operating in open systems where their interaction produces complex and non-linear outcomes. This ontological claim distinguished Bhaskar’s realism from two dominant alternatives. Against empiricism, Bhaskar argued that the regularities scientists observe are not the basic material of science — they are the effects of causal mechanisms, and those mechanisms exist independently of whether any regularity is observed. Against idealism, Bhaskar argued that the structures science investigates are not constructions of the human mind — they are real features of a world that exists independently of our knowledge of it.
Bhaskar’s (1979) subsequent work, The Possibility of Naturalism, extended this framework to social science, where it encountered the additional complexity that social structures are not independent of human activity in the way that natural structures are. Physical causal mechanisms operate whether or not anyone knows about them. Social causal mechanisms — including the psychological empowerment mechanism that is central to the CCF — require human activity for their production and reproduction, but are not reducible to that activity. The mechanism exists in the structured causal powers of the social position (the artisan’s role, their relationship to craft identity, their institutional context) rather than in any individual’s intentions or beliefs.
Three of Bhaskar’s foundational distinctions are essential for the CCF’s epistemological positioning. The first is the distinction between the real, the actual, and the empirical. The real consists of structures and mechanisms that exist independently of whether they are activated. The actual consists of events that occur when mechanisms are activated. The empirical consists of events that are observed and recorded. A psychological empowerment mechanism exists in the real — it is a structured causal power that artisans carry by virtue of their psychology, their work context, and their relationship to the studio’s social structure. Whether it is activated — whether it actually produces discretionary effort — depends on whether the enabling conditions are present. Whether it is observed — whether a measurement instrument captures its activation — depends on the instrument’s capacity to reach the relevant dimensions of artisan experience.
This three-level distinction is critical for understanding why the CCF’s effect sizes are bounded at ρ = .44 rather than approaching 1.0. The empowerment mechanism is real, but it operates in an open system where multiple other mechanisms — status dynamics, resource constraints, client pressure, seasonal production cycles — are simultaneously active. The meta-analytic correlation captures the average activation rate of the mechanism across diverse samples, not its theoretical maximum. A studio that implements the CCF correctly is not aiming to capture the full theoretical power of the empowerment mechanism. It is aiming to create the conditions under which the mechanism is activated more reliably and more completely than it was before implementation. That is what the 19.4% variance explained represents: not the ceiling of what empowerment can do, but the average documented effect when the enabling conditions are partially but not fully present.
Bhaskar’s second critical distinction is between closed and open systems. Natural science typically works in closed laboratory systems where mechanisms can be isolated, their operation stabilised, and their effects measured without interference from competing mechanisms. This is why physics produces the kind of precise, universal laws that management science cannot replicate. Social phenomena occur in open systems where multiple mechanisms operate simultaneously, where their interaction is non-linear, and where the conditions of their activation are themselves subject to change through the very process of investigation. The CCF’s supermodularity claim is an open system claim: it proposes that three mechanisms operating simultaneously produce outcomes that exceed the sum of their individual effects, in a context where the interaction is shaped by cultural conditions, macroeconomic constraints, and the specific history of each studio’s implementation. This claim cannot be tested in a closed system and should not be evaluated by closed-system standards. Pawson and Tilley (1997), in their Realistic Evaluation framework, proposed the appropriate standard: what mechanisms, activated in what contexts, produce what outcomes? The CCF’s five propositions are MCO specifications, and they should be evaluated as such.
Bhaskar’s third distinction — between generative mechanisms and empirical regularities — addresses the question of what kind of thing the CCF’s constructs are. An empiricist would say that Proximal Interference is a name for an empirical regularity: when founders are physically present in production environments, artisan decision-making rates are lower. A critical realist says that Proximal Interference names a generative mechanism: a real relational dynamic produced by the structural conditions of founder co-location, status differential, and evaluation apprehension, which has the causal power to suppress artisan discretionary effort regardless of whether any particular instance of that suppression is observed. The mechanism’s existence does not depend on consistent empirical observation. Its existence is a claim about the real — about the causal structures of craft enterprise social organisation — that is subject to empirical test but not defined by it.
Margaret Archer’s (1995) morphogenetic framework adds the temporal dimension that Bhaskar’s structural account requires. Archer argued that social structures exist prior to and independently of the individuals who reproduce or transform them — they are the inherited context within which individuals act — but that individual action can modify those structures over time, creating new structural conditions for subsequent action. The CCF’s founder identity transition is a morphogenetic process: the founder inherits the structural position of master craftsman (a position with specific causal powers and liabilities), acts within that position in ways that reproduce or modify its structural properties (the Safety Pre-Flight and Identity Bridge protocols are modification attempts), and in doing so creates new structural conditions (a studio with a different social organisation of craft knowledge) for the artisan workforce to act within. Andrew Sayer (1992, 2000) provided the methodological implications: social science should focus on the identification of causal mechanisms and the specification of the contextual conditions under which they operate, rather than on the statistical estimation of empirical regularities across large samples. This is the research programme the CCF’s doctoral article specified.
The CCF’s Ontological Commitments
The CCF makes three foundational ontological commitments — claims about what actually exists in the social world — that must be stated precisely because they determine the evidential standards to which the framework’s claims are subject.
The first commitment is that psychological empowerment is a real causal structure. The CCF claims that Spreitzer’s (1995) four-dimension construct describes a real feature of artisan psychology — a structured set of cognitive assessments (meaning, competence, self-determination, impact) that exists in the artisan’s psychological reality and has genuine causal powers: specifically, the power to produce or withhold the discretionary effort that transforms craft competence into craft excellence. This is a realist commitment: the construct is not merely a useful conceptual tool for organising observations. It is a description of something that actually exists and that has effects whether or not it is measured. The Seibert, Wang, and Courtright (2011) meta-analysis — 142 independent samples, ρ = .44 — provides confirmatory evidence for this reality. It does not define the reality. The distinction matters because it determines what it would mean to disconfirm the construct: the construct is disconfirmed not when a specific sample fails to reproduce the correlation but when the theoretical account of what produces the correlation is shown to be incorrect.
The second commitment is that organisational routines are real social structures with causal powers. The CCF claims that the habit-based routines of Pillar 2 are not merely patterns of behaviour — the empiricist account — but real social structures that have the causal power to coordinate action without requiring deliberate decision-making, to retain organisational memory beyond the individuals who established the routine, and to resist change through a structural inertia that is not reducible to individual resistance. Feldman and Pentland (2003) reconceptualised routines as generative systems with both ostensive and performative dimensions, arguing that routines are not static templates but dynamic systems that produce and reproduce social patterns through the recursive interaction of structure and action. This is a critical realist account of routine: the routine is a real social structure (the ostensive aspect) that is continuously reproduced through individual action (the performative aspect), and that has causal powers (coordination, memory retention, inertia) that are properties of the structure rather than of any individual participant’s behaviour.
The third and most philosophically ambitious commitment is that the Verified Empowered Routine is an emergent real property. Mario Bunge (1979), in his systematic ontology, defined emergence as the appearance of novel properties at higher levels of organisation that are not possessed by and cannot be predicted from the properties of the components at lower levels. The VER is an emergentist claim: the CCF asserts that when psychological empowerment, habit-based routines, and blockchain-verified institutionalisation are simultaneously operational, an organisational property comes into existence — the VER — that has causal powers (producing supermodular performance gains, sustaining craft identity investment, creating institutional memory) that none of the individual components possesses. Dave Elder-Vass (2010), in his careful reconstruction of emergentist ontology for social science, provided the critical qualification: emergence is real but not mysterious. Emergent properties are produced by the specific relational organisation of the components — by the way they interact, not merely by their co-presence. The VER’s emergence depends on the specific interaction structure the CCF specifies: the meaning activation that blockchain verification produces when empowered artisans experience the record as theirs, the cognitive bandwidth freed by routine automaticity that allows empowered initiative to express itself consistently, the accountability loop that blockchain documentation creates for routine compliance. These are the interaction conditions. Without them, three independently operating mechanisms do not produce the emergent property — they produce additive effects at best.
The constructivist challenge to these ontological commitments must be engaged directly. A social constructivist would argue that the CCF’s constructs — Proximal Interference, Architectural Professionalisation, the VER — are not descriptions of real social structures but conceptual tools that organise observations in theoretically useful ways. The constructs do not describe social reality; they constitute a way of seeing social reality that is useful for some purposes and not others. The critical realist response, drawing on Sayer (1992), is not to deny that conceptual tools shape what can be observed. It is to insist that the CCF’s constructs make claims that can be confirmed or disconfirmed — that Proximal Interference either suppresses artisan discretionary effort when the founder is physically present or it does not, and that this is a fact about the social world, not merely a consequence of adopting a particular conceptual framework. The constructivist is right that the CCF’s concepts are historically and culturally situated. The critical realist insists that situated concepts can nonetheless describe real mechanisms — and that the distinction between a useful concept and a correct one is preserved by the empirical programme.
Retroduction as the CCF’s Primary Inference Method
Charles Sanders Peirce (1878, 1903) identified three modes of scientific inference: deduction (deriving specific conclusions from general premises), induction (deriving general conclusions from specific observations), and abduction — the inference from surprising observations to the best available explanation of what could have produced them. Peirce regarded abduction as the logic of discovery: the mode of inference through which genuinely new theoretical ideas are generated, as distinct from the testing of existing ideas (deduction) and the accumulation of evidence for them (induction). Brian Haig (2005), extending Peirce’s framework for psychological research methodology, proposed an abductive theory of scientific method in which theory generation through abduction precedes and is logically distinct from theory testing through deduction and induction.
Critical realism extends this framework through what Danermark, Ekström, Jakobsen, and Karlsson (2002) called retroduction — the specifically critical realist mode of inference that moves from observed phenomena to the hypothetical causal structures that could have produced them. Retroduction asks: given that this outcome has occurred, what must be true about the causal structures of the world for this outcome to be possible? Retroduction is not hypothesis generation in the positivist sense — it does not merely propose associations to be tested. It proposes specific causal structures (mechanisms, conditions, relations) whose existence would make the observed phenomenon intelligible, and it does so at the level of the real rather than the empirical.
The CCF’s theory-building is retroductive in this precise sense. The starting observation is a phenomenon that management science had not adequately theorised: founder-dependent craft micro-enterprises with demonstrably capable artisan workforces consistently fail to grow beyond the founder’s personal productive capacity, despite the founder’s evident desire to grow and the artisans’ evident craft competence. This phenomenon is surprising — if capability and motivation are present, why does performance not scale? The retroductive question is: what must be true about the causal structures of this organisational context for this outcome to be possible? The CCF’s answer — Proximal Interference operating through status dynamics and evaluation apprehension, founder identity investment in the master craftsman role, tacit knowledge concentrated in a single individual — is a retroductive inference about the real structures that produce the observed phenomenon. These structures were not observed directly. They were inferred as the most coherent account of what must be true for the observations to make sense.
The contrast with the positivist approach illuminates what is distinctive about the CCF’s epistemology. A positivist researcher would approach the craft enterprise growth problem by testing whether specific interventions — empowerment programmes, documented procedures, technology adoption — correlate with revenue growth across a sample of studios. Finding a significant correlation would constitute evidence for the intervention’s effectiveness. Finding no correlation would constitute evidence against it. The CCF’s retroductive approach is different: it begins not from the correlation between intervention and outcome but from the mechanism that produces the correlation. It asks why empowerment produces growth when it does, what structural conditions allow the mechanism to be activated, and what happens when those conditions are absent. Pawson and Tilley (1997) argued that this is the only approach that produces actionable knowledge for practitioners: knowing that an intervention correlates with outcomes across a sample does not tell a practitioner whether the intervention will work in their specific context, because the correlation is the product of mechanisms operating under specific contextual conditions that may or may not be present in the practitioner’s studio.
The abduction-retroduction cycle is the epistemological engine through which the CCF generates new knowledge rather than merely accumulating confirming evidence. Abduction generates the initial hypothesis: the surprising observation that capability without growth is a consistent pattern in craft micro-enterprises leads to the abductive inference that a structural mechanism must be suppressing the translation of capability into performance. Retroduction refines the hypothesis: if Proximal Interference is the mechanism, what specific structural conditions produce it, what observable consequences does it have, and what interventions would address the conditions rather than the symptoms? The empirical programme tests the refined hypothesis against observations. Anomalies — cases where the predicted pattern does not obtain — generate new retroductive inferences that further specify the mechanism’s scope conditions and interaction effects. This is the cycle through which the CCF research programme advances.
The Mechanism-Context-Outcome Framework
Pawson and Tilley’s (1997) realistic evaluation framework — the Mechanism-Context-Outcome configuration — provides the most appropriate evaluative standard for the CCF’s claims. The central insight of realistic evaluation is that interventions do not produce outcomes directly. Mechanisms produce outcomes when they are activated by appropriate contexts. The same mechanism can produce different outcomes in different contexts, and the same context can produce different outcomes when different mechanisms are operating. This framework resolves the generalisability problem: the CCF does not claim to produce the same outcomes in all contexts. It claims that specific mechanisms produce specific outcomes when specific contextual conditions are present. The five MCO configurations correspond to the CCF’s five theoretical propositions.
P1 — The Empowerment Mechanism. Mechanism: psychological empowerment activates the structured causal power of artisan craft identity, producing discretionary effort that exceeds what compliance-based management can generate. Context: high-skill, low-volume craft production environments where artisans have been selected for demonstrated competence, where tacit knowledge application is the primary performance driver, and where the enabling conditions of psychological safety and founder identity transition are present. Outcome: measurable improvement in task performance (Seibert et al. 2011, ρ = .44), tacit knowledge transfer (Argote & Ingram 2000), and craft quality at the ceiling of the artisan’s developed competence. Disconfirmation condition: a well-powered study of craft micro-enterprises in which psychological empowerment scores do not predict task performance above the baseline, controlling for prior performance and ability, would require revision of the mechanism’s scope conditions or its specification in high-craft contexts.
P2 — The Routine Mechanism. Mechanism: habit-based operational routines free the cognitive bandwidth that would otherwise be consumed by coordination decisions, creating the structural opportunity for empowered artisan initiative to express itself consistently rather than variably. Context: founder-dependent studios where operational knowledge is currently concentrated in the founder’s personal judgment, where the routines are introduced with enabling rather than coercive framing (Adler & Borys 1996), and where habit formation has reached sufficient automaticity (Lally et al. 2010, mean 66 days). Outcome: reduction in founder-dependency for operational decisions, consistent quality execution across the artisan workforce, and institutional memory of operational knowledge that persists beyond individual tenure. Disconfirmation condition: a longitudinal study finding that studios with formal operational routines do not show reduced founder decision-escalation rates relative to studios without routines, controlling for studio age and size, would require revision of the opportunity mechanism specification.
P3 — The Verification Mechanism. Mechanism: blockchain-verified provenance activates the meaning dimension of psychological empowerment when the verification system is designed with Transparency Equity principles — specifically, when the artisan experiences the record as their portable professional identity asset rather than the studio’s surveillance infrastructure. Context: artisan workforce in which craft identity is partially constituted by the visibility and verification of craft decisions, the Digital Inclusion Protocol has ensured equitable access regardless of digital literacy, and the Week 9 meaning conversation has framed the Passport correctly. Outcome: enhanced work engagement through meaning activation (Jiang & Cheng 2025), provenance premium at the 15–25% floor estimate from cross-category meta-analysis, and institutional memory of craft reasoning. Disconfirmation condition: a controlled study finding that Transparency Equity-designed Digital Passport implementation does not increase artisan scores on the meaning dimension of Spreitzer’s (1995) empowerment scale — relative to a control condition — would require revision of the meaning activation mechanism.
P4 — The Supermodularity Mechanism. Mechanism: the three pillars are Edgeworth complements — each increases the marginal return of the others — producing the emergent VER whose causal powers exceed the sum of the individual pillars. Context: full three-pillar deployment with enabling formalisation, Transparency Equity design, and the five prerequisites met. The interaction conditions specified by Elder-Vass (2010) must be present: verification changes what routines mean to the artisan (meaning activation pathway), routines free the cognitive bandwidth for empowered initiative (opportunity pathway), and empowerment reduces the incentive to falsify records (integrity pathway). Outcome: statistically significant positive β₄ in the three-way interaction regression, above main effects and two-way interactions, with adequate power as specified by Aguinis and Vandenberg (2014). Disconfirmation condition: a well-powered regression study (N ≥ 30 studios, N ≥ 300 artisans) finding a non-significant β₄ for the three-way interaction of empowerment, routines, and digital verification would require revision of the supermodularity claim, likely toward the additivity account that Gerhart (2007) proposes as the alternative.
P5 — The Partial Adoption Mechanism. Mechanism: because the three pillars are complements rather than substitutes, adopting fewer than three produces a performance valley in which transition costs are incurred without synergistic returns. Context: any studio implementing one or two pillars without the third. Outcome: performance worse than either the prior state or the full three-pillar system, for the duration of partial implementation. Disconfirmation condition: a longitudinal study finding that two-pillar studios achieve equivalent or superior performance outcomes to three-pillar studios over a 36-month horizon would falsify P5 and would require fundamental revision of the complementarity architecture.
Pawson (2006) and Wong, Greenhalgh, Westhorp, Buckingham, and Pawson (2013) developed the realistic synthesis methodology for evidence accumulation in mechanism-based research programmes. Rather than pooling effect sizes across studies — the meta-analytic standard appropriate for positivist research programmes — realistic synthesis asks: across the studies that have tested this mechanism in this context, what is the most coherent account of the conditions under which the mechanism is activated and the outcomes it produces? The CCF’s evidence programme should produce realistic syntheses rather than conventional meta-analyses for its central claims — not because meta-analysis is inappropriate, but because the mechanism-context-outcome framework requires a different kind of evidence integration.
The Stratified Ontology of Craft Enterprise Transformation
Bhaskar’s (1975) stratified ontology proposes that reality is composed of multiple levels, each with emergent properties and causal mechanisms that are neither reducible to nor explicable solely in terms of lower levels. Elder-Vass (2010) extended this framework for social science, arguing that social structures exist at multiple levels of organisation — from the neural to the institutional — and that causal mechanisms at each level can produce effects at other levels through upward and downward causation. The CCF’s interventions and outcomes are distributed across four levels of this stratification, and understanding the level-specific causal mechanisms is essential for understanding why the framework produces the effects it claims and why those effects are bounded by the conditions at other levels.
At the first, neurological and cognitive level, Wood and Neal’s (2007) habit automaticity operates as a feature of the artisan’s neural architecture — the shift of behavioural control from the prefrontal cortex (deliberate decision-making) to the basal ganglia (pattern-triggered automaticity) that Lally et al. (2010) measured across the 18–254 day formation range. Spreitzer’s (1995) four empowerment dimensions exist at this level as cognitive assessments — structured evaluations of meaning, competence, self-determination, and impact that are properties of the artisan’s psychology and that have causal effects on behaviour through their influence on motivation, effort allocation, and discretionary engagement. This is the level at which the empowerment intervention directly operates.
At the second, interpersonal level, Edmondson’s (1999) psychological safety operates as a shared property of the team — a collective belief that is produced through the interaction of individual assessments and that has causal powers at the team level that no individual’s assessment produces alone. Proximal Interference operates at this level as a relational dynamic between the founder and the artisan workforce — produced by their specific spatial, status, and evaluative relationship, and having causal effects on the artisan’s individual-level empowerment assessments through the mechanism of evaluation apprehension. The Safety Pre-Flight protocol intervenes at this level by modifying the relational dynamic through the founder’s Active Error Modelling behaviour. Kozlowski and Klein’s (2000) multilevel theory specifies the aggregation conditions under which individual-level empowerment assessments constitute a team-level climate — the conditions that must be met before the CCF’s individual-level interventions can be said to have produced organisational-level effects.
At the third, organisational level, the VER exists as an emergent real property — produced by the specific relational organisation of the three pillars but not reducible to any of them. Zucker’s (1986) crystallised institutionalisation operates at this level as the process through which the VER becomes self-sustaining — no longer requiring active maintenance by the founder but reproduced through the taken-for-granted practices of a workforce that has internalised the CCF’s norms as their professional identity. The Digital Passport functions at this level as an organisational artefact in Feldman and Pentland’s (2003) sense — a material element of the organisational routine that anchors the routine’s ostensive specification in a form that resists performative drift. The CCF’s Pillar 2 and Pillar 3 interventions operate primarily at this level.
At the fourth, market level, Spence’s (1973) signalling mechanism operates between the studio and luxury consumers — a market-level causal structure in which verified quality signals are interpreted by buyers to reduce the information asymmetry that Akerlof (1970) identified as the structural problem in experience goods markets. The provenance premium is a market-level outcome produced by the interaction of the studio’s signalling behaviour and the buyer’s interpretive framework. Vigneron and Johnson’s (1999) consumer segmentation specifies the contextual conditions at this level: the premium mechanism operates for quality-driven buyers whose purchase decision is responsive to quality verification, not for status-driven buyers whose purchase decision is responsive to social signalling that verification may undermine.
The CCF’s interventions work at Levels 2 and 3 to produce outcomes at Levels 3 and 4. The causal pathway crosses levels — individual-level empowerment (Level 1) produces team-level safety climate (Level 2) that enables organisational-level routine compliance (Level 3) that produces market-level provenance premium (Level 4). This cross-level causal pathway is what makes the CCF’s claims non-trivial: it is not sufficient to produce empowered individuals. The empowerment must be produced in conditions that allow it to constitute a team-level climate, which must be embedded in an organisational structure that converts it into consistent craft performance, which must be verified through a mechanism that communicates its quality to the market. Each transition between levels requires specific conditions, and the failure of any transition suppresses the outcome at all subsequent levels.
The South African macroeconomic constraints that Feng (2025) documents operate at a fifth level — the societal and institutional level — that can suppress outcomes at all lower levels regardless of the quality of Level 2 and Level 3 interventions. Load-shedding disrupts the digital infrastructure on which Level 3 routines depend. Rand volatility creates Level 4 pricing uncertainty that compresses the provenance premium. POPIA compliance costs create Level 3 governance burdens that constrain the Hybrid Architecture’s implementation. These are not failures of the CCF’s mechanism. They are demonstrations of the open system character of social causation: mechanisms that operate at lower levels are subject to suppression and amplification by conditions at higher levels.
Theory Progress Criteria
Imre Lakatos (1978) proposed that scientific theories should be evaluated not as isolated hypotheses but as research programmes — structured sets of assumptions, auxiliary hypotheses, and empirical commitments that evolve over time through the progressive or degenerative modification of their peripheral elements. A progressive research programme is one whose modifications generate novel predictions that are subsequently confirmed. A degenerative programme is one whose modifications are post-hoc adjustments to protect the core from disconfirmation without generating new predictive content. Alan Chalmers (1999) extended Lakatos’ framework to address the philosophy of science questions that arise when evaluating theoretical contributions in contested disciplines — which management science certainly is.
The CCF’s hard core consists of three claims that are definitional to the framework. First, that psychological empowerment, habit-based externalisation, and blockchain-verified institutionalisation are complementary mechanisms that produce supermodular gains when deployed together in founder-dependent craft micro-enterprises. Second, that Proximal Interference is a real relational dynamic — produced by founder co-location, status differential, and evaluation apprehension — that systematically suppresses empowerment activation in the target context. Third, that the VER is a real emergent property of the three-pillar system, with causal powers that none of the individual pillars possesses. These claims cannot be abandoned without abandoning the CCF as a distinctive theoretical framework.
Four criteria define theoretical progress in the CCF research programme.
Progress Criterion 1 is novel empirical confirmation of the Proximal Interference construct. The construct advances the research programme when it is tested through the within-subject design specified in the doctoral article — measuring artisan decision-making rates, self-determination scores, and creative performance under conditions of founder presence versus absence, controlling for task complexity and prior performance. A significant effect in the predicted direction confirms the construct as a real mechanism with the causal properties the CCF claims. A null effect requires revision: either the causal mechanism is not as specified (evaluation apprehension may not be the primary pathway), or the scope conditions are more restrictive than proposed (the effect may require specific cultural or relational conditions that are not universal to craft micro-enterprises), or the measurement operationalisation does not adequately capture the decision-making suppression the mechanism is claimed to produce. A null effect is not a fatal disconfirmation of the CCF’s core claims — it is a mandate for mechanism refinement.
Progress Criterion 2 is boundary condition specification through cultural moderator research. The CCF’s cultural moderator gap — the acknowledged absence of evidence for the framework’s effectiveness in high-power-distance, collectivist South African studio contexts — advances when empirical research identifies the specific conditions under which the empowerment mechanism is and is not activated across the cultural diversity of the South African jewellery workforce. This is a progressive modification: not a defence of the core against disconfirmation, but a refinement of the scope conditions that generates novel predictions about which studios in which cultural contexts will respond most strongly to Pillar 1 interventions.
Progress Criterion 3 is mechanism pathway differentiation for the VER’s meaning activation claim. The VER advances when the specific pathway through which blockchain verification activates empowerment’s meaning dimension is empirically traced and differentiated from alternative pathways. Jiang and Cheng’s (2025) finding provides the initial evidence. A progressive research programme would design studies that distinguish between the identity investment pathway (the verification activates meaning because it constitutes the artisan’s professional identity), the institutional memory pathway (the verification activates meaning because it creates a permanent record of craft decisions), and the social recognition pathway (the verification activates meaning because it makes the artisan’s contribution visible to the market). Each pathway makes different predictions about which artisan populations will respond most strongly to which Digital Passport design features. Differentiating the pathways is more informative than confirming the aggregate effect.
Progress Criterion 4 is rival theory engagement with the dynamic capabilities account. Teece, Pisano, and Shuen’s (1997) dynamic capabilities framework is the strongest rival account of craft enterprise growth: it would explain the CCF’s observed outcomes as the product of the founder’s sensing, seizing, and reconfiguring capabilities rather than as the product of the specific three-pillar complementarity architecture. The CCF research programme advances when it designs a comparative study that distinguishes the dynamic capabilities account from the Architectural Professionalisation account — specifically, a study that measures both dynamic managerial capabilities and CCF implementation fidelity in a sample of craft micro-enterprises, and tests which construct better predicts the performance outcomes the CCF claims. If dynamic capabilities predicts better, the CCF must engage the dynamic capabilities literature more centrally. If CCF fidelity predicts better, the Architectural Professionalisation construct has demonstrated its surplus value over the existing theoretical alternatives.
The Abduction-Retroduction Cycle in Practice
The abduction-retroduction cycle is not merely an epistemological description of how the CCF was built. It is an ongoing methodological commitment about how the research programme should proceed. Each round of structured devil’s advocate interrogation that the CCF has undergone across twelve rounds of Librarian research games is an abductive exercise: the Librarian identifies surprising observations — cases where the CCF’s predicted mechanisms should not operate, anomalies in the evidence base, contexts where the scope conditions are not met — and the CCF’s theorists generate retroductive inferences about what causal structures would need to be modified to account for the anomalies.
The Digital Exclusion anomaly is an example of this cycle in practice. The CCF’s initial formulation of the Transparency Equity principle assumed universal digital literacy — a scope condition that was not explicitly specified because it was not recognised as a scope condition. The anomaly was identified through the Librarian’s Round 12 interrogation: what happens when a master artisan cannot independently access the Digital Passport? The abductive inference was that the Transparency Equity mechanism could produce the opposite of its intended effect — creating a two-tier artisan class rather than equalising access to provenance documentation. The retroductive inference was that the scope condition of digital literacy had been implicit in the mechanism’s specification and needed to be made explicit. The Digital Inclusion Protocol is the structural response to the retroductive inference — a modification of the auxiliary hypothesis (Passport implementation) that protects the hard core (Transparency Equity) while addressing the anomaly.
The Scale-Up Paradox is a second example. The initial CCF formulation did not specify an upper boundary on its scope conditions — it addressed the 8–25 person range without specifying what happens when successful implementation produces growth beyond that range. The abductive inference from the anomaly (successful CCF studios growing toward the Phase 2 threshold) was that the safety climate mechanism, which the CCF had theorised as operating through the founder’s direct Active Error Modelling, would face activation failure as the studio departmentalised and the founder’s direct reach contracted. The retroductive inference was that the mechanism requires replication at the department level — a Safety Climate Cascade — rather than operating solely through the founder. The Phase 2 Architecture is the structural response.
Each anomaly-driven modification of the CCF’s auxiliary hypotheses generates new empirical predictions that were not present before the modification. This is the Lakatosian criterion for a progressive research programme: the modifications are not merely defensive but generative. A research programme that adds new scope conditions without adding new predictions is degenerative. A research programme whose scope condition refinements generate novel, testable predictions is progressive. The CCF’s trajectory — from its initial three-pillar formulation through twelve rounds of critical interrogation to its current architecture of prerequisites, growth protocols, digital inclusion provisions, aesthetic protection mechanisms, and apprenticeship architecture — has been consistently progressive in this sense: each modification has generated new empirical predictions while maintaining the integrity of the core three-pillar complementarity claim.
The Philosophical Status of the CCF
The CCF is a critical realist research programme with a mechanism-context-outcome evaluative framework, a stratified ontology that locates its interventions and outcomes at specific levels of social reality, and progress criteria that define what genuine theoretical advancement looks like. It is not a positivist prediction machine — it does not claim that implementation produces determinate outcomes regardless of context. The bounded effect sizes, the scope conditions, and the mechanism-context specifications are not rhetorical hedges. They are direct implications of the critical realist commitment to the open-system character of social causation.
It is not an interpretivist framework. The CCF insists that the causal mechanisms it identifies — Proximal Interference, empowerment, routine institutionalisation — are real structural features of craft enterprise social organisation, not merely useful ways of interpreting founder and artisan experience. The founder’s meaning-making is causally relevant — it determines whether the Safety Pre-Flight produces genuine safety or compliance theatre — but meaning-making does not constitute the mechanism. It activates or suppresses it. This is the critical realist distinction between the subjective conditions of mechanism activation and the objective reality of the mechanism itself.
It is a critical realist framework that takes seriously both the causal powers of social structures and the meaning-constituted nature of human action — and that proposes interventions at the level where those two dimensions of social reality intersect. The organisational level is where empowered artisans executing verified routines produce a social structure — the VER — that has real causal properties that neither the artisans nor the routines nor the verification system possesses in isolation. That is the philosophical claim. The empirical programme specified across ten articles, five theoretical propositions, and four progress criteria is the test. The test has not yet been fully conducted. The framework is ready for it.
Frequently Asked Questions
1. Why critical realism rather than pragmatism, which has a stronger tradition in management research and would accommodate the CCF’s combination of positivist and interpretivist elements?
Pragmatism — associated in management research with scholars like Richard Rorty and, in a different register, with the action research tradition — would accommodate the CCF’s method pluralism by treating theories as tools that are evaluated by their practical usefulness rather than their correspondence to mind-independent reality. This is a coherent philosophical position, and it would resolve the epistemological tensions the CCF faces by dissolving rather than answering them: there are no tensions between positivist and interpretivist methods if all methods are evaluated purely by their practical utility. The reason the CCF is not a pragmatist framework is that it makes claims that cannot be evaluated solely by practical utility. The claim that Proximal Interference is a real relational dynamic that suppresses artisan discretionary effort is not merely a useful way of organising management interventions. It is a claim about a real causal structure that either exists or does not exist, and whose existence has implications beyond the practical context of any specific studio. The critical realist commitment to the reality of causal mechanisms — and to the empirical programme that tests their reality — is not merely philosophically preferable to pragmatism. It is required by the CCF’s specific combination of theoretical ambition and empirical accountability. Merton (1968) positioned middle-range theory as the appropriate aspiration for management research precisely because it makes claims that are neither so abstract as to be untestable nor so concrete as to be merely descriptive. The CCF’s critical realist foundation gives those claims a philosophical account of why they are simultaneously bounded in their scope and real in their referents.
2. The CCF’s critical realist foundation commits it to claims about what really exists in social reality. How does the framework handle the acknowledged fact that social structures change over time and that the mechanisms it identifies may not be stable across historical periods?
Archer’s (1995) morphogenetic framework provides the theoretical response. Social structures are real but not eternal — they are produced and reproduced through human action, and they can be modified through that action over time. The empowerment mechanism that Spreitzer (1995) identified operates through the specific social structures of contemporary craft production: the artisan workforce’s relationship to craft identity, the founder’s social position as master craftsman, the luxury consumer’s interpretive framework for quality signals. These structures are historically contingent — they emerged in specific social conditions and they will change as those conditions change. The CCF’s claim is not that the empowerment mechanism is universally and timelessly real. It is that the mechanism is real in the specific structural conditions of contemporary South African craft enterprise production — conditions that have a history and will have a future. This is consistent with critical realism’s commitment to ontological depth (the mechanism is real) and historical contingency (the conditions of its activation are historically specific). The research programme’s boundary condition work — identifying the cultural, institutional, and macroeconomic conditions under which the mechanisms are activated — is precisely the empirical engagement with historical contingency that critical realism requires.
3. The MCO configurations for P4 and P5 specify disconfirmation conditions that could falsify the CCF’s core claims. Has the CCF considered the reputational risk of publishing these conditions publicly?
Yes, and the decision to publish them is deliberate. The CCF’s entire credibility architecture is built on the Spence (1973) costly signalling principle: the willingness to disclose limitations and disconfirmation conditions is itself a signal of theoretical seriousness that competitors who do not name their falsification conditions cannot replicate. A management framework that does not specify what evidence would disconfirm it is not a scientific framework — it is a belief system. The reputational risk of naming falsification conditions is real but it is the appropriate cost of claiming theoretical credibility at the level this series aspires to. If the three-way interaction regression produces a non-significant β₄ in a well-powered study, the CCF must revise its supermodularity claim. That is the appropriate consequence. The alternative — claiming supermodularity without specifying what evidence would falsify it — would be exactly the kind of empirically irresponsible theoretical overclaiming that the CCF’s approach to intellectual honesty has consistently rejected. Lakatos (1978) was precise about this: a research programme that cannot specify what would constitute disconfirmation of its core claims is not a progressive programme. It is a degenerative one in which the theory is insulated from its evidence base by protective rhetorical hedges rather than by genuine auxiliary hypothesis specification. The CCF is a progressive programme. It earns that status by naming what would disconfirm it.
4. How does the CCF’s stratified ontology handle the relationship between the individual artisan level and the organisational level — specifically, can individual artisan empowerment be said to constitute a studio-level property without the aggregation fallacy that Chen et al. (2007) identified?
Kozlowski and Klein’s (2000) multilevel theory provides the technical answer. Individual empowerment scores constitute a studio-level climate only when the aggregation conditions are met: within-group agreement (rwg ≥ .70) must be demonstrated before individual scores are averaged to the studio level. Studios where empowerment is highly variable across team members — where senior setters score high and junior polishers score low, as in the composite case — do not have a studio-level empowerment climate in Kozlowski and Klein’s sense. They have a two-tier distribution that aggregation would misrepresent. The critical realist response adds a layer to this technical answer. The studio-level empowerment climate that Kozlowski and Klein require for valid aggregation is, in critical realist terms, an emergent property — a real structural feature of the studio’s social organisation that comes into existence when the individual assessments are sufficiently convergent, and that has causal powers (team coordination, collective initiative, shared craft identity) that the individual assessments do not produce in isolation. The dual-level measurement instrument in the CCF’s implementation protocol — Spreitzer’s (1995) individual scale plus Kirkman and Rosen’s (1999) team empowerment scale — is precisely the response to this ontological requirement: the individual scores and the team-level construct are measured separately, not derived from each other, because they exist at different levels of the stratified ontology and have different causal powers at those levels.
5. The CCF’s abduction-retroduction cycle is described as the epistemological engine of the research programme. But abduction produces the most plausible explanation, not necessarily the correct one. How does the CCF distinguish between abductive inferences that track real mechanisms and those that produce coherent but incorrect accounts of causal structure?
Peirce (1903) acknowledged that abduction produces plausible rather than certain conclusions, and that multiple abductive inferences may be equally consistent with the observations that prompted them. The CCF’s approach to this epistemological limitation operates through three practices. First, theoretical triangulation: the CCF’s abductive inferences are tested against multiple independent theoretical traditions — the empowerment mechanism is not only consistent with Spreitzer’s (1995) construct but with SDT’s intrinsic motivation framework (Deci & Ryan 2000), with Argote and Ingram’s (2000) knowledge transfer account, and with Edmondson’s (1999) psychological safety research. Convergence across independent traditions increases the probability that the abductive inference tracks a real mechanism rather than a theoretical artefact. Second, retroductive specificity: the more specifically a retroductive inference specifies the causal structure — its components, their interaction conditions, the outcomes they produce when activated — the more constraining it is empirically, and the less likely it is to be merely an artefact of the theorist’s conceptual framework. Proximal Interference’s specification — the ratio of bench-level decisions escalated to the founder, measured over two weeks, with the predicted founder-presence versus founder-absence comparison — is specific enough to be strongly disconfirmable. Third, anomaly productivity: a retroductive inference that tracks a real mechanism should be productive when it encounters anomalies — it should generate new predictions that resolve the anomalies rather than requiring post-hoc adjustment. The Digital Exclusion and Scale-Up Paradox modifications were productive: each generated new empirical predictions that the original CCF formulation could not have made. This productivity is the pragmatic indicator — in Haig’s (2005) sense — that the retroductive inferences are tracking real mechanisms rather than merely coherent constructions.
References
Ackroyd, S., & Fleetwood, S. (Eds.). (2000). Realist perspectives on management and organisations. Routledge.
Adler, P. S., & Borys, B. (1996). Two types of bureaucracy: Enabling and coercive. Administrative Science Quarterly, 41(1), 61–89.
Aguinis, H., & Vandenberg, R. J. (2014). An ounce of prevention is worth a pound of cure: Improving research quality before data collection. Annual Review of Organizational Psychology and Organizational Behavior, 1(1), 569–595.
Akerlof, G. A. (1970). The market for “lemons”: Quality uncertainty and the market mechanism. Quarterly Journal of Economics, 84(3), 488–500.
Archer, M. S. (1995). Realist social theory: The morphogenetic approach. Cambridge University Press.
Argote, L., & Ingram, P. (2000). Knowledge transfer: A basis for competitive advantage in firms. Organizational Behavior and Human Decision Processes, 82(1), 150–169.
Bhaskar, R. (1975). A realist theory of science. Leeds Books.
Bhaskar, R. (1979). The possibility of naturalism: A philosophical critique of the contemporary human sciences. Harvester Press.
Bunge, M. (1979). Treatise on basic philosophy: Vol. 4. Ontology II: A world of systems. Reidel.
Chalmers, A. F. (1999). What is this thing called science? (3rd ed.). University of Queensland Press.
Danermark, B., Ekström, M., Jakobsen, L., & Karlsson, J. C. (2002). Explaining society: Critical realism in the social sciences. Routledge.
Deci, E. L., & Ryan, R. M. (2000). The “what” and “why” of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11(4), 227–268.
Edmondson, A. C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383.
Elder-Vass, D. (2010). The causal power of social structures: Emergence, structure and agency. Cambridge University Press.
Feldman, M. S., & Pentland, B. T. (2003). Reconceptualizing organizational routines as a source of flexibility and change. Administrative Science Quarterly, 48(1), 94–118.
Feng, Y. (2025). Digital transformation and employees’ unethical pro-self behaviours: Evidence from South African manufacturing. SA Journal of Business Management.
Fleetwood, S., & Ackroyd, S. (Eds.). (2004). Critical realist applications in organisation and management studies. Routledge.
Gerhart, B. (2007). Modeling HRM and performance linkages. In P. Boxall, J. Purcell, & P. Wright (Eds.), Oxford handbook of human resource management (pp. 552–580). Oxford University Press.
Haig, B. D. (2005). An abductive theory of scientific method. Psychological Methods, 10(4), 371–388.
Ibarra, H. (1999). Provisional selves: Experimenting with image and identity in professional adaptation. Administrative Science Quarterly, 44(4), 764–791.
Jiang, H., & Cheng, X. (2025). Perceived algorithmic control and gig workers’ work engagement: Assessing the mediating role of psychological empowerment. Frontiers in Psychology. PMC.
Kozlowski, S. W. J., & Klein, K. J. (2000). A multilevel approach to theory and research in organizations. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel theory, research, and methods in organizations (pp. 3–90). Jossey-Bass.
Lakatos, I. (1978). The methodology of scientific research programmes. Cambridge University Press.
Lally, P., van Jaarsveld, C. H. M., Potts, H. W. W., & Wardle, J. (2010). How are habits formed: Modelling habit formation in the real world. European Journal of Social Psychology, 40(6), 998–1009.
Merton, R. K. (1968). Social theory and social structure. Free Press.
Milgrom, P., & Roberts, J. (1990). The economics of modern manufacturing: Technology, strategy, and organization. American Economic Review, 80(3), 511–528.
Milgrom, P., & Roberts, J. (1995). Complementarities and fit: Strategy, structure, and organizational change in manufacturing. Journal of Accounting and Economics, 19(2–3), 179–208.
Pawson, R. (2006). Evidence-based policy: A realist perspective. Sage.
Pawson, R., & Tilley, N. (1997). Realistic evaluation. Sage.
Peirce, C. S. (1878). Deduction, induction, and hypothesis. Popular Science Monthly, 13, 470–482.
Peirce, C. S. (1903). Harvard lectures on pragmatism. (Published in Collected Papers, Vol. 5, 1934.)
Repenning, N. P., & Sterman, J. D. (2002). Capability traps and self-confirming attribution errors in the dynamics of process improvement. Administrative Science Quarterly, 47(2), 265–295.
Sayer, A. (1992). Method in social science: A realist approach. Routledge.
Sayer, A. (2000). Realism and social science. Sage.
Seibert, S. E., Wang, G., & Courtright, S. H. (2011). Antecedents and consequences of psychological and team empowerment in organizations: A meta-analytic review. Journal of Applied Psychology, 96(5), 981–1003.
Spence, M. (1973). Job market signaling. Quarterly Journal of Economics, 87(3), 355–374.
Spreitzer, G. M. (1995). Psychological empowerment in the workplace: Dimensions, measurement, and validation. Academy of Management Journal, 38(5), 1442–1465.
Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18(7), 509–533.
Vigneron, F., & Johnson, L. W. (1999). A review and a conceptual framework of prestige-seeking consumer behavior. Academy of Marketing Science Review, 1(1), 1–15.
Wong, G., Greenhalgh, T., Westhorp, G., Buckingham, J., & Pawson, R. (2013). RAMESES publication standards: Realist syntheses. BMC Medicine, 11, 21.
Wood, W., & Neal, D. T. (2007). A new look at habits and the habit-goal interface. Psychological Review, 114(4), 843–863.
Zucker, L. G. (1986). Production of trust: Institutional sources of economic structure, 1840–1920. Research in Organizational Behavior, 8, 53–111.
