Building What Makes
Standards Matter
Understanding how standards acquire force, and building what can be built
Published, Sound, and Inert
The criterion was published. The failure happened anyway.
The Baltimore fire and the national hose thread standard that followed appear elsewhere on this site as the case for what standards deliver. The record continues past the point where that account stops.
The standard was published in 1905. Nine years later, 287 of the roughly 8,000 American cities and towns had conforming couplings and hydrants. The specification was public, technically correct, and endorsed by the national fire, insurance, and waterworks bodies, and nearly the entire country ignored it. Where adoption moved, it moved where a promise was at stake: cities that had signed mutual aid compacts, promising their neighbors help in a major fire, discovered that the promise was unkeepable across mismatched threads. The obligation converted the published specification into a live requirement. Where no such obligation existed, the standard went ignored. In 1991, 87 years after Baltimore, the Oakland firestorm killed 25 people while mutual aid crews struggled to connect to Oakland's hydrants. Oakland adopted the national standard afterward. Several major American cities use their own threads to this day.
The specification was never the problem. What was missing is the subject of this program.
The Infrastructure Problem
A correct, publicly available criterion can be inert for decades while the failures it addresses continue. The inertness has a structure of its own, and most of it goes unseen. Evidence of nonconformance arises and dies in an inbox. A finding that would matter to an insurer never reaches one. The parties who possess levers never learn that the criterion exists, and the parties who know the criterion possess no levers.
The people who could change this rarely see themselves as connected. An underwriter setting application questions, a procurement officer writing bid requirements, a lawyer arguing what a reasonable institution should have known, a journalist receiving five complaints that are one story, a professional body drafting guidance: each occupies a position in the same missing machinery, and almost none of them know the others are load-bearing. The condition extends well beyond technical standards. Researchers and practitioners in evidence-based fields, from public health to housing policy, describe that experience from the knowledge side: well-supported findings that never become decision-relevant to the institutions positioned to act on them.
A published criterion is an answer. The substrate is what makes anyone have to ask.
Realis builds standards for institutional decision-making, and this problem governs whether any such standard, ours or anyone's, ever matters. So the lab has opened a research program on the question directly: what structural conditions allow a publicly specified criterion to acquire distributed practical force without centralized enforcement, and under what conditions does an excellent criterion die on the shelf?
The Candidate Architecture
The program begins with a six-part candidate architecture: six functions that may have to be present for a published criterion to acquire practical force. The research is designed to determine whether those functions are real, whether all six are necessary, and whether they relate in the way the candidate model proposes. The object under study is the condition under which the whole arrangement works, above any one component.
Legibility. An observer can recognize that a condition described by the criterion may exist, without mastering the full apparatus.
Record. The recognition becomes a durable artifact that survives the observer.
Portability. The artifact is intelligible across organizational and professional boundaries.
Interest. A receiving actor has an existing exposure the evidence touches: liability, underwriting risk, fiduciary duty, procurement eligibility, professional obligation, statutory oversight.
Consequence. The receiver possesses a lever: premiums, coverage terms, contract conditions, discovery, audit findings, licensing, publication, purchasing.
Feedback. Anticipated consequence alters upstream behavior before failure, and the criterion becomes decision-relevant in advance.
This is a candidate, and the design says so. It has not been tested, and a model like this one earns little by fitting successful cases after the fact: once a criterion has propagated, it is too easy to reconstruct a plausible pathway in hindsight. The program puts its evidential weight on the failure side: historical cases where the same criterion propagated in one place and died in another, coded against pre-registered predictions about which links were absent, with a stopping rule frozen before scoring that specifies the result under which the model is wrong as such. The metric system in the United States, legally authorized since 1866 and inert in daily commerce ever since, is in the case bank because it falsifies the assumption that authority produces adoption. If the model fails, the failure will be published under the same documentation discipline the corpus asks of the institutions it studies.
What the Program Is Doing
The research is built on matched historical pairs: the same criterion, propagated in one arm and inert in the other, so the specification is constant while the surrounding structure varies. Hose threads in mutual-aid corridors against the holdout cities; international paper sizing worldwide against North America; attestation regimes that procurement demanded against near-identical siblings no one asked for. Alongside the coding, the program builds and tests experimental apparatus in live channels: a question set an underwriter could pilot, a disclosure mechanism a procurement office could trial, a transmission pathway counsel could evaluate. These prototypes are labeled experimental everywhere they appear, and a prototype is experimental apparatus however widely it is deployed or praised; what it produces is data.
Historical coding can only look backward, where every success explains itself. A prototype placed in a real channel is the one way to observe a criterion acquiring force forward in time.
Who This Needs
The lab cannot do this alone, and should not. The problem is closer to designing for flight than designing an engine. Expertise already exists in the component systems: underwriters know underwriting; lawyers know the pathways by which evidence acquires legal consequence; procurement specialists know their own purchasing constraints; researchers know how findings are produced and evaluated. The question at this level is different: what functions must the whole system contain for a good criterion to become consequential, how do those functions connect, and what happens when one is missing?
The people who occupy these channels understand things about them that no outside coding can recover. The program is built so that knowledge shapes the machinery while it is being worked out, and can overturn it.
If you set underwriting questions, write procurement requirements, argue reasonableness, receive the complaints, draft professional guidance, oversee accreditation, regulate, investigate, or research adjacent questions: part of this may be yours to supply. Collaborators can attack the candidate architecture, supply cases and prior art the program has missed, name the constraints in their own channels, and prototype components inside them. There are legitimate professional, research, and commercial opportunities in what gets built, and the program says so plainly.
No one is being asked to adopt a Realis standard. A procurement officer can require disclosure of decision conditions without requiring anyone's certification; a lawyer can introduce published criteria into a reasonableness argument without any institution declaring anything; an underwriter can add three questions to a form. The components work severally.
The structure also requires no center. The insurer and the procurement office may never learn the other exists, and the architecture under investigation is one in which it works anyway. Realis is building no coalition. The lab's role is to characterize the machinery and publish enough of the design that people who already possess its component capabilities can build independently.
Nothing here is a result yet. The commitments are.
The full research architecture, including the case bank, the coding scheme, the pre-registered predictions, and the stopping rule, is available to collaborators.
The program's commitments are published in full before any results exist: How Published Criteria Acquire Force: A Candidate Architecture and Pre-Registered Research Design, a pre-results working paper recording the hypothesis, method, case bank, predictions, and stopping logic, with its results sections gated until the research earns them.
Contact Realis Institute to challenge the model, contribute a case, or prototype a component in your channel.
Neighboring Work
The questions next door.
Everett Rogers · The adopters
That a good idea does not spread on its merits is the founding result of a field. Everett Rogers built diffusion research on it: the spread of an innovation is a communication process, conducted through channels, over time, among the members of a social system, and adoption is a decision made under uncertainty in which most people look to others like themselves who have already adopted. The metric system is one of his examples of an innovation that takes decades. His account reaches well past voluntary individual choice: he distinguishes optional, collective, authority, and contingent innovation-decisions, and he shows a criterion's decision type being changed by law, as when seat belts moved from an owner's option to a legislated requirement. The rate of spread, the shape of the adoption curve, the characteristics of early and late adopters, and the attributes of an innovation that make it easier to take up are all charted ground.
In his terms, the mutual aid compact can be read as changing the hose thread from an optional decision into a contingent one, and his taxonomy can classify that change. This page tests something smaller: what produced that change. The six functions describe a criterion becoming consequential through invocation: the underwriter who adds three questions to a form, the procurement office that writes a disclosure requirement, the counsel who introduces a published specification into an argument about what a reasonable institution should have known. None of them implements the criterion as the practice being changed. Whether that invocation is itself an adoption event in his sense, or whether the operative unit is the pathway by which a criterion acquires force in a decision made elsewhere, is one of the things the case coding is meant to settle.
Joseph Farrell & Garth Saloner, Paul David · The switch
The hose-thread record has an economics. Joseph Farrell and Garth Saloner showed that an industry already standardized on one arrangement can lock in even when a better one is available, a result they called excess inertia, and that under incomplete information the inertia is a coordination failure that communication among the parties can reduce. Paul David supplied a related historical account of path dependence, in which technical interrelatedness, scale economies, and quasi-irreversible investment can make early contingencies increasingly consequential. His example, the keyboard, has been argued over since; the mechanism is what he contributed, and it does not depend on the example.
Through their model, every city with a local thread was standardized, and the national specification asked 8,000 of them to switch. The mutual aid compact looks like the communication mechanism their model calls for, made binding: a promise to a neighbor that could not be kept across mismatched couplings. It fits, and the program counts it as a hypothesis. The matched pairs are built to test it. Each pair fixes the specification and the coordination problem as constant as the record permits while comparing the surrounding structure, so that what the program then asks the six-function model to explain in the inert arm is not the economics of switching, which both arms share, but which function was absent when the standard died on the shelf.
Implementation science · The uptake
The experience described above from the knowledge side, well-supported findings that never become decision-relevant to the institutions positioned to act on them, has a field of its own. When the journal Implementation Science launched in 2006, Martin Eccles and Brian Mittman defined implementation research as the scientific study of methods that promote the systematic uptake of research findings and other evidence-based practices into routine practice. Their scope already included professional and organisational behaviour and clinical, community, and policy contexts, and the field since has consolidated the conditions that determine whether evidence reaches routine use into frameworks that span the intervention itself, the outer setting of external policy, incentives, and peer pressure, the inner setting of organisational culture and leadership, the individuals involved, and the process of putting a change into effect.
Those surrounding levers exist and are well mapped. The object under test is whether a published criterion can become consequential through a pathway whose carrier never implements the criterion as the practice being changed: the underwriter's question set, the procurement office's disclosure clause, the reasonableness argument in court. The live-channel prototypes place criteria into those channels to learn whether that is a separable mechanism or an implementation strategy the field has already described. The laboratory would rather learn the answer from the field's own work than assert it.
Everett M. Rogers, Diffusion of Innovations (Free Press, 1962; 5th edn 2003).
Joseph Farrell and Garth Saloner, "Standardization, Compatibility, and Innovation," RAND Journal of Economics 16, no. 1 (1985), 70–83.
Paul A. David, "Clio and the Economics of QWERTY," American Economic Review 75, no. 2 (1985), 332–337.
Martin P. Eccles and Brian S. Mittman, "Welcome to Implementation Science," Implementation Science 1, no. 1 (2006).
Laura J. Damschroder, David C. Aron, Rosalind E. Keith, Susan R. Kirsh, Jeffery A. Alexander, and Julie C. Lowery, "Fostering Implementation of Health Services Research Findings into Practice: A Consolidated Framework for Advancing Implementation Science," Implementation Science 4, no. 50 (2009).
Banner: general arrangement and instrument panel drawings of the North American X-15, from the aircraft's utility flight manual, T.O. 1X-15-1, United States Air Force, public domain. The X-15 was a research aircraft: it existed because some questions could not be answered without flying something. It was built to put hypotheses into a machine, instrument the result, and revise. That is this program's method. The plate is a cutaway: the finished airplane hides nearly everything required to make it fly, and opening it up exposes the systems and their relationships. That is this program's subject. And the drawing's own title names the object of study. The component disciplines already exist; what the program is trying to discover is the arrangement in which they collectively produce the property the whole system is for. The shortest description of that research is printed on the plate: the general arrangement.