Contribution
It names the step between an individual intervention and an organizational change as a separate problem, propagation, and gives it two conditions: the leverage of the members whose flexibility rose and the value content of the culture. It recasts Bond’s organizational flexibility model (the “orgflex”, Bond 2018) as one case of that mechanism, the case where the people who change are the leaders. It is a theory paper: the authors call the propositions “hypotheses, not a set of findings” and say the two measures a test needs do not yet exist.
Key insight
Flexibility does not add up. A raised repertoire in one member persists and recurs in untrained colleagues only if that member actually changes the reinforcement and punishment those colleagues meet, and only inside a culture that does not penalize the action. The mediator to be measured is a change in consequences, not a change in what members report.
Assumptions
- Functional contextualism. Behaviour is act-in-context; psychological flexibility is a repertoire of six processes (the hexaflex), not a trait, and fusion is a position on a continuum set by context.
- Multilevel selection. Through the Prosocial framework (Atkins, Wilson and Hayes 2019), with Ostrom’s core design principles as generalized by Wilson, Ostrom and Cox (2013), selection by consequences is applied to the interlocking action of many people.
- Culture as a contingency structure. Culture is “the consequences members deliver to one another, together with the rules and shared statements of purpose that modulate their function”. §5.3 calls this “a deliberate simplification”.
- Values versus goals. In the ACT sense values are inexhaustible directions and goals exhaustible targets; profit is treated as a goal.
- Cross-level parallel. Defusion changes the function of a private event at Level 1; culture changes the function of an adverse consequence at Level 3. The paper says the parallel is functional, that “no proposition rests on it”, and that it is “assumed rather than demonstrated”.
Key results
There are no empirical results. The propositions (§4):
- P1. Individual flexibility is positively associated with worker health. Stated as the established premise, not a contribution, and in §5.3 as resting “largely” on cross-sectional evidence.
- P2. Applying the Prosocial method turns individually held values into jointly held ones, measured by a self-report index and a behavioural convergence index scored against target repertoires taken from a source independent of the trained group. The one proposition that claims the effect of a specific intervention.
- P3. A value-based culture moderates the flexibility–health link. What moderates is value content, not resources or job control; it is falsified if a high-control, low-value-content setting converts flexibility into health as fully as a value-based one. It is set against Bond and Flaxman (2006), where job control moderates.
- P4 (the central one). Retention and spread of a rise in individual flexibility depend jointly on leverage and culture. Leverage is “effective control over the consequences others encounter, not position on the organizational chart”, vertical (appraisals, resources, tasks) or horizontal (peer accountability, Ostrom’s principle 3 and graduated sanctions). Falsified if training low-leverage members in a penalizing culture yields action that is retained and taken up by others. Leverage must be measured a priori, “through an index of who controls whose consequences”, or the proposition becomes circular.
- P5. The resulting flexibility shows as group-level functioning (purpose-consistent action under pressure, adaptation that keeps purpose, coordination that survives disruption). Not yet testable: the measure must be built first and kept independent of P4’s.
- Design (§5.1). Leverage high/low crossed with culture value-based/extractive, in three studies: a negative control, leverage varied with culture held, culture varied with leverage held.
- Diagnostic (§3.3, §5.2). Compliance and values-consistent action look alike while members are watched; they separate when oversight relaxes.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Individual-level wellbeing interventions do not reliably change organizations | weak | Fleming (2024), cross-sectional, which by the paper’s own account measured participants’ wellbeing, not organizations |
| C2 | Propagation of individual flexibility depends jointly on leverage and culture (P4) | weak | conceptual argument; stated as a hypothesis with a falsifier |
| C3 | Bond’s leader route is a special case of the leverage condition | moderate | definitional, given the paper’s definition of leverage |
| C4 | Compliance and value-based culture separate when surveillance drops | weak | argument; no study cited |
| C5 | Value content, not job control, moderates flexibility’s effect on health (P3) | weak | hypothesis, set against Bond and Flaxman (2006) |
Concepts
- Propagation: “the retention over time of a values-consistent repertoire in those who acquired it, under conditions in which that repertoire carries a local cost, together with its recurrence in colleagues who were not themselves trained”.
- Leverage: effective control over the consequences colleagues meet; vertical or horizontal.
- Organizational flexibility: a property of how members’ actions interlock (Glenn’s interlocking behavioural contingencies), not a state of any member nor a property over and above them.
- Value-based culture: one whose shared purpose changes what reinforces members, so that action holds without surveillance.
- Niche construction: an organization’s insulation of itself from extractive market selection by governance, ownership or market segment; “insulation of degree, not immunity”.
Connections
Its individual-scale premise is ACT’s psychological flexibility (Hayes et al. 2006), for which the record holds Kashdan and Rottenberg (LIT-553) and THEORY-050. Its organizational antecedent is Bond (2018), a chapter in Wilson, Hayes and Biglan (eds.), Evolution and Contextual Behavioral Science, pp. 131–148, which the record has not registered. Its collective level is the Prosocial framework and Ostrom’s principles. Its account of culture is Glenn’s (2004) cultural selection. It cites no work on opinion dynamics or network influence, though its leverage index is a directed graph of who sets whose consequences.
Bearing on the record
- CLAIM-349. The claim’s note says Bond’s model “could not be identified uniquely”. The paper identifies it: Bond, F. W. (2018), “Organizational flexibility: creating a mindful and purpose-driven organization”. The paper agrees with CLAIM-349 that flexibility is context-sensitive and that more control is not more flexibility, and its compliance-versus-value test is close to the claim’s RO-DBT translation (less disclosure under tighter control). It does not separate access to a signal from its authority over action; its unit is a repertoire and the consequences it meets, not a signal. So it is a prior transfer of ACT to the organization, as the claim already says, but not a prior statement of the claim’s three-way split.
- The influence model. P4’s leverage is close to what the record’s influence canvas (CLAIM-297, CLAIM-282) computes: who weights whose output. The paper offers no dynamics for it. A reduction of P4 to a weighted influence network would be the record’s contribution, and the paper’s demand that leverage be measured independently of spread is the same circularity guard QUESTION-117 needs.
No instruction for machine-learning practice; nothing for the anthology.
Limitations
- No data, and no measure. The two constructs on which every test depends, an a priori leverage index and a group-level index of systemic functioning, do not exist; the authors say so.
- The motivating evidence is thinner than stated. Fleming (2024) is cross-sectional and, as the paper describes it, found that participants in individual-level interventions were “no better off” than non-participants on wellbeing. That bears against individual benefit, the half of the premise the paper takes as settled, as much as against organizational transfer, which it did not measure. Calling it “direct evidence” that individual interventions “do not carry through to the organization” overstates it.
- Errors and slips. Two empirical sentences in §2.1 (“Systematic reviews and meta-analyses report consistent reductions …” and “A large-scale study across three Scandinavian countries …”) carry no citation. Bakker and Demerouti’s “Job demands–resources theory” is listed as 2016 with volume 22 of the Journal of Occupational Health Psychology, which is the 2017 volume (not checked against the journal). In the PDF the commentary under Proposition 2 is laid out after the heading of Proposition 4.
- Normative stance. Worker health is treated as a value in its own right throughout; the authors say so, and it shapes P3 and §5.2.
Open questions
- Can leverage be measured before and independently of spread, as P4 demands? A network map of who sets whose appraisals, resources and assignments is proposed but not built.
- Does P4 reduce to a known model of weighted social influence, in which case its predictions follow from the network alone?
Note of 2026-10-29: Bond’s chapter, registered, and its citation corrected
Connections gives Bond (2018) as a chapter in Wilson, Hayes and Biglan (eds.), pp. 131–148, not registered. It is now registered as LIT-1830, Unavailable, and the citation is wrong in two places. Checked against the book’s table of contents, it is chapter 11, pp. 185–199, of D. S. Wilson and S. C. Hayes (eds.), Evolution and Contextual Behavioral Science (Context Press, 2018). Anthony Biglan wrote the foreword and is not an editor. The same correction is recorded on LIT-1218.